## 1.  INTRODUCTION

## Source details

**Canonical URL:** [1.  INTRODUCTION](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022012-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022012-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022012-print-pdf.pdf.json)

---

### Research objective and scope
- Aim: study effects of different types of COVID-19 containment measures on infection cases, economic activity, and fiscal space (sovereign risk).
- Sample: daily data from February 26,2020 to June 30,2021 covering a set of 44 advanced and emerging economies.
- Frequency and method: high-frequency daily data; local projection á la Jordà(2005) building on Deb et al.(2020a).
- Key proxies:
  - Economic activity proxied by NO2 emissions.
  - Fiscal space proxied by sovereign 5-year CDS spreads.
  - Containment measures drawn from Oxford’s Coronavirus Government Response Tracker (OxCGRT) with three indices ranging from "physical" closures to "smart" measures.
- Figure notes:
  - Cross-country average OxCGRT "stringency index" normalized between[0, 1] on the right y-axis.
  - Cross-country average CDS spreads measured in basis points on the left axis.
  - Infection (shaded) waves shown are those in the Western Hemisphere.
  - Argentina is excluded from the computation of the cross-country average of CDS spreads.

### Conceptual framing
- Fiscal space interpreted along two dimensions: long-term sustainability and market access/financing.
- CDS spreads used as an indicator for fiscal space (market access/perceived sovereign risk), consistent with Kose et al.(2017).
- A priori ambiguity: early pandemic uncertainty may have driven CDS spreads independently of initial policy interventions.

### Principal empirical findings (headline)
- Trade-offs between infections and economic activity:
  - Containment measures can limit infection cases but potentially at the expense of economic activity.
  - Degree of trade-off depends on containment type:
    - "Physical" measures (e.g., lockdowns, school closings, travel bans) are most effective in containing outbreaks but most disruptive to economic activity.
    - "Smart" measures (e.g., contact tracing, mask mandates) can contain infections to some degree while safeguarding the economy.
- Fiscal space / sovereign risk:
  - Smart measures may be as effective as physical ones in improving sovereign risk (reducing 5-year CDS spreads).
  - Containing infection cases via stricter containment is associated with lower CDS spreads; smart measures show reductions in CDS spreads quantitatively comparable to physical measures.
- Dynamics across waves:
  - During the first wave, tighter containment and CDS spreads moved together; in subsequent waves the relationship appeared to reverse, possibly because financial markets later internalized policy measures.
- State- and condition-dependent findings:
  - Stricter physical containment measures tend to affect economic activity less and reduce CDS spreads more in EMs versus AEs.
  - Faster public health response time can more quickly normalize economic activity after an initial shock and improve the sovereign risk profile.
  - Stronger initial public finances (low public debt) are associated with a better economic outlook and lower sovereign risk when containment measures are implemented.

### Fiscal announcements and sovereign risk (EU-19 focus)
- New database: constructed a new database of "daily" fiscal announcements in response to COVID-19 for EU-19 countries (extended compared to prior work).
- Empirical approach: state-dependent local projection; interpret fiscal announcements as unanticipated fiscal shocks á la Ramey (2011b,a), building on Deb et al. (2021).
- Result: sovereign risk is improved under a combination of large support packages and smart measures.

---

### 3.  DATA

### Dataset overview and key variables
- Daily-frequency panel for 44 countries spanning from February 26 2020 until June 30 2021.
- COVID-19 variables: infection cases, deaths, tests, vaccinations.
- Containment indices: physical, physical and smart, smart (all normalized between 0 and 1).
- Economic activity indicators: NO2 emissions (daily, city-level), temperature, humidity, pm10.
- Financial variables: sovereign yields and CDS spreads at maturities 1, 3, 5 and 10 years (5-year CDS used as benchmark).
- Fiscal policy announcements at daily frequency for EU-19 countries (expanded and recategorized).

### Data sources
- COVID-19 cases, deaths, tests, vaccinations: Johns Hopkins University Coronavirus Resource Center.
- Containment measures: Oxford’s Coronavirus Government Response Tracker (OxCGRT).
- NO2, temperature, humidity, pm10: Air Quality Open Data Platform of the World Air Quality Index (WAQI); meteorological stations report data three times per day and the median reported values are used.
- Sovereign yields and CDS spreads: Refinitiv Eikon Datastream and Bloomberg.
- Fiscal announcements: Yale’s COVID-19 Financial Response Tracker (CFRT), cross-checked with IMF Policy Tracker and other sources; European institutions’ announcements included where not reported at country level.

### Containment indices (OxCGRT-based)
- Physical (stringency) index: physical closure policies across eight dimensions (school closing, workplace closing, public events cancelling, restrictions on gatherings, closing public transport, stay at home requirements, restrictions on internal movement, international travel controls).
- Containment and health index ("physical and smart"): expands physical index by including testing policy, contact tracing, facial coverings, vaccination policy.
- Smart index: includes only smart containment measures.
- All three indices normalized between 0 and 1 so a unitary impulse shock equals a 0 to 100 percent increase in the policy index (e.g., a unitary shock in the physical index equals full lockdown).

### Economic activity indicator validation (NO2)
- Panel regression (monthly frequency) of ∆log(IPI) on ∆log(NO2) shows:
  - ∆log(NO2) coefficient = 0.137 ** (t-statistic 2.67)
  - Constant = 0.00362 *** (t-statistic 4.25)
- Note: NO2 emissions measured in parts per billion (ppb).
- Interpretation: a 1 percent change in NO2 implies a 0.137 percent variation in the Industrial Production Index (IPI).

### Fiscal announcements and categorization
- Fiscal measures recorded as percentage of national (2020 and 2021) GDP.
- Macro-categories: emergency lifeline and demand-support, plus health and non-health measures.
  - Lifeline measures: cash-flow support to firms and households (e.g., credit guarantees, loans).
  - Demand-support: measures increasing income of firms and households (e.g., wage subsidies, tax payment forbearances).
- Fiscal announcements treated as fiscal news to identify unanticipated fiscal policy shocks.

### Empirical strategy (summary)
- Estimation: Linear panel Local Projections (LPs) following Jordà (2005).
  - LP specification: y_{i,t+h} = α_i + β^h η_{i,t} + Ψ^h X_{i,t} + Φ^h(L) y_{i,t} + ε_{i,t+h}, h = 0, 1, 2, ..., H.
  - Inference with country-clustered standard errors.
- Projection horizon: 30 days (1 month); confidence intervals at 90 percent.
- Main dependent variables (in log deviations from t−1): COVID-19 cases, NO2 emissions, 5-year sovereign CDS spreads.
- Shock variables: physical index, physical and smart index, smart index.
- Controls include:
  - COVID-19 deaths, vaccinations, vaccination policy.
  - Temperature, humidity, pm10.
  - CDS spreads at 1, 3, 5, 10 years (in logs).
  - Sovereign yields at 1 and 10 years.
  - When dependent variable is economic activity or fiscal space, include COVID-19 cases as a control.
- Robustness: VIX included in robustness checks (results do not vary when VIX is included).
- Identification mitigations:
  - Interact physical index with a seasonality factor to account for seasonal variation and voluntary social distancing.
  - Include lags of dependent variable and country-specific time trends.
- State-dependent LPs: use dummy indicator D_t to separate regimes and study heterogeneity (AEs vs EMs, fast vs slow PHRT, high vs low public debt).

### Baseline empirical results (quantitative headline effects)
- General trade-off:
  - More physical lockdowns → lower COVID-19 cases but larger negative impact on economic activity.
  - Impacts strongest for physical lockdowns versus physical and smart, and smart-only measures.
- Quantitative headline effects (from unitary shocks in indices; indices scaled 0–1):
  - COVID-19 cases:
    - A unitary shock in the "physical" and "physical and smart" indices is associated with a 90 percent drop in COVID-19 cases at peak (stronger under physical).
    - A unitary shock in the "smart" index is associated with a 20 percent drop in COVID-19 cases at peak.
  - Economic activity (NO2 emissions):
    - A unitary shock in the "physical" index is associated with a peak 90 percent drop in NO2 emissions after a few days.
    - A unitary shock in the "physical and smart" index yields a milder plunge in NO2 emissions compared to physical-only.
    - A unitary shock in the "smart" index yields no statistically significant drop in NO2 emissions.
  - Fiscal space (5-year CDS spreads):
    - CDS spreads drop by about 20 percent on average across the three types of containment measures; IRFs are consistently negative throughout the projection horizon.
    - This corresponds to approximately a 25 basis points drop in CDS spreads in the sample.
- Interpretation:
  - Smart measures soften the health-economy trade-off: milder reductions in COVID-19 cases, preserve economic activity, and deliver statistically comparable reductions in sovereign risk relative to physical lockdowns.

---

### APPENDIX B.  ADDITIONAL RESULTS (FIGURES AND TABLES) — Baseline results and state-dependencies

### Baseline IRF presentation conventions
- Responses reported in log-deviations from past values.
- Shaded green areas represent the 90 percent confidence intervals; black solid lines represent median IRFs.

### Baseline findings (recap)
- Containment measures limit infection cases but can reduce economic activity.
- Physical lockdowns effective in reducing infection cases but disruptive to the economy.
- Smart measures can control infections to some degree while safeguarding economic activity.
- Smart measures are statistically as effective as physical ones in improving sovereign risk (reducing 5-year CDS spreads).

### State-dependencies — Advanced versus Emerging economies
- AEs typically have lower perceived sovereign risk than EMs.
- Tightening of the physical index implies a stronger drop in NO2 emissions under the AE state, peaking at -100 percent.
- CDS spreads under the AE regime tend to drop less than under the EM regime; in the EM regime spreads decrease consistently throughout the projection horizon.
- Quantitative note: "That is equal to a 30 percent reduction of CDS spreads on average across the projection horizon."
- Interpretation: financial markets may reward EMs more for taking difficult policy actions and/or sovereign risk in AEs is affected more by fundamentals.

### State-dependencies — Fast versus Slow public health response time (PHRT)
- PHRT (baseline): days to reach maximum stringency after 100 COVID-19 cases; state dummy = 1 when PHRT above median (slow), 0 when below median (fast).
- Findings:
  - Tightening of the physical index associated with an initial drop in NO2 emissions in both regimes; shock dissipates faster under the fast PHRT regime.
  - CDS spreads are statistically reduced only under the fast PHRT regime.
- Robustness: alternative PHRT definitions considered (e.g., 250 cases per 100 thousands inhabitants; days to reach a 1 percent tightening after 100 cases).

### State-dependencies — High versus Low public debt
- High-debt definitions:
  - (i) debt-to-GDP above 90 percent cutoff;
  - (ii) observations above the 75th percentile of the public debt sample distribution.
- Findings:
  - 5-year CDS spreads decrease significantly throughout the projection horizon only when public debt is low.
  - Peak response about -20 percent around 20 days after the shock (overall sample).
  - Considering EMs only, the decrease in CDS spreads peaks to -50 percent at the end of the projection horizon when public debt is low.
  - Figure 24 reports that the cut in NO2 emissions is more subdued under low public debt.
- Implication: containment measures can be more effective in reducing sovereign risk when initial fiscal conditions are stronger.

---

### APPENDIX A.  FISCAL ANNOUNCEMENTS DATASET — construction, classification, and measurement

### Dataset construction and coverage
- Starting point: Yale’s COVID-19 Financial Response Tracker (CFRT); methodology follows Deb et al. (2021).
- Supplementary sources: IMF Policy Tracker and other reports for cross-checking.
- Country coverage focus: "EU-19".
- EU-level fiscal measures distributed to each country by each country’s GDP shares when measures implemented at the EU level.
- Time coverage: "Feb2020 to end-June2021" (daily).

### Classification and aggregation
- Policy-instrument categorization and macro-aggregation follow IMF’s Fiscal Monitor database.
- Macro-categories: Lifeline and demand (spending and revenue) support; above-the-line and below-the-line measures; health and non-health measures.

### Definitions
- Lifeline measures include liquidity injections, loans, umbrella guarantees, credit guarantees, equity injections, asset purchases, targeted loans.
- Demand-support measures include wage subsidies, targeted transfers, grants, unemployment benefits, wage supplements, deferrals of tax and social security contributions, tax relief, support to SMEs.
- Above-the-line measures include unemployment benefits, grants and transfers, tax cuts and relief, payment deferrals, payment forbearances, grants to SMEs.
- Below-the-line measures include loans, capital injections, asset purchases, guarantees, government guarantees to banks/firms/households, equity injections.

### Measurement and interpretation
- Fiscal announcements reported as the size of the shock in "percent of national GDP".
- Series interpreted as unanticipated fiscal news/shocks as in Ramey (2011b,a).
- Figures referenced for Austria illustrate raw CFRT data, the expanded dataset after cross-checking and categorization, and categorical breakdowns.
- IMF’s Fiscal Monitor database access reference preserved exactly: "IMF’s    Fiscal    Monitor    database    can    be    accessed    at    https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19"

---

### APPENDIX B.  ADDITIONAL RESULTS — Fiscal announcements and state-dependent fiscal transmission

### Fiscal announcements — linear LP estimates (EU-19)
- Linear local projection (eq. 1) — effect of fiscal announcement (FPA) shock on CDS and NO2 on impact and after 4 weeks (Table2):
  - CDS on impact: 0.0378 (t-statistic (0.93))
  - CDS after 4 weeks: 0.223 ∗∗∗ (t-statistic (3.18))
  - NO2 on impact: -0.201 (t-statistic (-1.14))
  - NO2 after 4 weeks: 2.070 ∗∗ (t-statistic (2.48))
  - Significance codes: * p<0.1, ** p<0.05, *** p<0.01
- Interpretation:
  - On average, an increase in fiscal support is associated with a statistically significant mild rise in CDS spreads 4 weeks after the shock.
  - The same shock yields a doubling in NO2 emissions 4 weeks after the impulse shock.
  - Implication: fiscal policy interventions can support the economy but at the cost of slightly worsening sovereign risk.

### Interaction of fiscal announcements and containment measures
- Size of fiscal announcements did not matter for the transmission of physical index shocks on CDS spreads (median effects negative).
- Fiscal space improves when a mix of "large" fiscal support and "smart" measures are in place.
- Interpretation: large support packages may be expected to worsen public finances and therefore future default risks embedded in 5-year CDS spreads; however, when combined with smart containment measures that rely on testing and avoid physical lockdowns, the confidence boost and more positive outlook can reduce expected sovereign risk.

### Concluding remarks (appendices synthesis)
- Data and methods:
  - Daily data from February 2020 to June 2021 for 44 advanced and emerging economies.
  - Proxies: OxCGRT indices for containment measures, NO2 emissions for economic activity, CDS spreads for fiscal space.
  - Econometric approach: local projection à la Jordà (2005) and specification building on Deb et al. (2020a).
- Key summary findings:
  - Smart containment measures are relatively optimal: can contain infections, avoid disruptions to economic activity, and improve sovereign risk.
  - State-dependent evidence:
    - In EMs versus AEs, stricter physical containment measures tend to affect economic activity less and reduce CDS spreads more.
    - Fast versus slow PHRT: faster response helps economic activity normalize more quickly and improves sovereign risk profile.
    - Low versus high initial public debt: containment measures associated with improved economic outlook and lower sovereign risk when public debt is low.
  - Sovereign risk improves under a combination of large fiscal support and smart measures.
- Future research: further exploit and expand the fiscal announcement dataset to gauge effects of fiscal news on fiscal space and economic activity during COVID-19.

*Source: IMF working paper section "1. INTRODUCTION", "3. DATA", Appendices A and B, and REFERENCES (wpiea2022012-print-pdf).*

### 1.  INTRODUCTION

### 1.  INTRODUCTION

### Research objective and scope
- Aim: study effects of different types of COVID-19 containment measures on infection cases, economic activity, and fiscal space (sovereign risk).
- Sample: daily data from February 26,2020 to June 30,2021 covering a set of 44 advanced and emerging economies.
- Frequency and method: high-frequency daily data; local projection á la Jordà(2005) building on Deb et al.(2020a).
- Key proxies:
  - Economic activity proxied by NO2 emissions.
  - Fiscal space proxied by sovereign 5-year CDS spreads.
  - Containment measures drawn from Oxford’s Coronavirus Government Response Tracker (OxCGRT) with three indices ranging from "physical" closures to "smart" measures.
- Figure notes:
  - Cross-country average OxCGRT "stringency index" normalized between[0, 1] on the right y-axis.
  - Cross-country average CDS spreads measured in basis points on the left axis.
  - Infection (shaded) waves shown are those in the Western Hemisphere.
  - Argentina is excluded from the computation of the cross-country average of CDS spreads.

### Key conceptual framing
- Fiscal space is multifaceted, covering mainly two dimensions: long-term sustainability and market access/financing.
- In this paper, CDS spreads are used as an indicator for fiscal space (market access/perceived sovereign risk), consistent with Kose et al.(2017) and related literature.
- The relationship between containment measures and fiscal space is a priori ambiguous; early pandemic uncertainty may have driven CDS spreads independently of initial policy interventions.

### Principal empirical findings
- Trade-offs between infections and economic activity:
  - Baseline results suggest containment measures can limit infection cases but potentially at the expense of economic activity.
  - The degree of trade-off depends on the type of containment measure:
    - "Physical" measures (e.g., lockdowns, school closings, travel bans) can be most effective in containing outbreaks but are most disruptive to economic activity.
    - "Smart" measures (e.g., contact tracing, mask mandates) can contain infections to some degree while safeguarding the economy.
- Fiscal space / sovereign risk:
  - Smart measures may be as effective as physical ones in improving sovereign risk (reducing 5-year CDS spreads).
  - Containing infection cases via stricter containment is associated with lower CDS spreads; smart measures show reductions in CDS spreads quantitatively comparable to physical measures.
- Dynamics across waves:
  - During the first wave, tighter containment and CDS spreads moved together; in subsequent waves the relationship appeared to reverse, possibly because financial markets later internalized policy measures.
- State-dependent and conditional findings:
  - Stricter physical containment measures tend to affect economic activity less and reduce CDS spreads more in EMs versus AEs.
  - Faster public health response time can more quickly normalize economic activity after an initial shock and improve the sovereign risk profile.
  - Stronger initial public finances (low public debt) are associated with a better economic outlook and lower sovereign risk when containment measures are implemented.

### Fiscal announcements and sovereign risk (EU-19 focus)
- New database: constructed a new database of "daily" fiscal announcements in response to COVID-19 for EU-19 countries (extended compared to prior work).
- Empirical approach: state-dependent local projection; interpret fiscal announcements as unanticipated fiscal shocks á la Ramey (2011b,a), building on Deb et al. (2021).
- Result: sovereign risk is improved under a combination of large support packages and smart measures.

### Methodological and data notes
- The paper follows Deb et al.(2020a) and leverages daily-frequency indicators (NO2, CDS, OxCGRT indices).
- Uses three different containment measures indices from OxCGRT that span from "physical" closures to "smart" measures.
- Justification for CDS proxy: exclusive reliance on high-frequency data prevents use of government budget or debt series to proxy fiscal space; CDS spreads are the best high-frequency proxy available.
- Comparison to prior literature:
  - Extends time coverage relative to some earlier studies (e.g., Cevik and Ozturkkal (2020)) by using data through end-June 2021 and a wider set of controls.
  - Confirms and refines findings that smart/fast containment measures can reduce infections while safeguarding economic resources (Hosny (2021); Fotiou and Lagerborg (2021); Deb et al. (2020a)).

*Source: IMF working paper section "1. INTRODUCTION" (wpiea2022012-print-pdf).*

### 3.  DATA

### 3.  DATA

### Dataset overview
- Daily-frequency panel for 44 countries spanning from February 26 2020 until June 30 2021.
- Variables included:
  - COVID-19-related variables: infection cases, deaths, tests, vaccinations.
  - Containment measures (indices): physical, physical and smart, smart (all normalized between 0 and 1).
  - Economic activity indicators: NO2 emissions (daily, city-level), temperature, humidity, pm10.
  - Financial variables: sovereign yields and CDS spreads at maturities 1, 3, 5 and 10 years (5-year CDS used as benchmark).
  - Fiscal policy announcements at daily frequency for EU-19 countries (expanded and recategorized).

### Data sources
- COVID-19 cases, deaths, tests, vaccinations: Johns Hopkins University Coronavirus Resource Center.
- Containment measures: Oxford’s Coronavirus Government Response Tracker (OxCGRT).
- NO2, temperature, humidity, pm10: Air Quality Open Data Platform of the World Air Quality Index (WAQI); meteorological stations report data three times per day and the median reported values are used.
- Sovereign yields and CDS spreads: Refinitiv Eikon Datastream and Bloomberg.
- Fiscal announcements: Yale’s COVID-19 Financial Response Tracker (CFRT), cross-checked with IMF Policy Tracker and other sources; European institutions’ announcements included where not reported at country level.

### Containment indices (OxCGRT-based)
- Physical (stringency) index: physical closure policies across eight dimensions (school closing, workplace closing, public events cancelling, restrictions on gatherings, closing public transport, stay at home requirements, restrictions on internal movement, international travel controls).
- Containment and health index (named "physical and smart"): expands physical index by including testing policy, contact tracing, facial coverings, vaccination policy.
- Smart index: includes only smart containment measures.
- All three indices are normalized to be between 0 and 1 so that a unitary impulse shock equals a 0 to 100 percent increase in the policy index (e.g., a unitary shock in the physical index equals full lockdown).

### Economic activity indicator validation (NO2)
- Panel regression (monthly frequency) of ∆log(IPI) on ∆log(NO2) shows:
  - ∆log(NO2) coefficient = 0.137 ** (t-statistic 2.67)
  - Constant = 0.00362 *** (t-statistic 4.25)
  - Note: NO2 emissions are measured in parts per billion (ppb).
- Interpretation: a 1 percent change in NO2 implies a 0.137 percent variation in the Industrial Production Index (IPI).

### Fiscal announcements and categorization
- Fiscal measures recorded as percentage of national (2020 and 2021) GDP.
- Policy instruments recategorized into macro-categories: emergency lifeline and demand-support, plus health and non-health measures.
  - Lifeline measures: cash-flow support to firms and households (e.g., credit guarantees, loans).
  - Demand-support: measures increasing income of firms and households (e.g., wage subsidies, tax payment forbearances).
- Fiscal announcements treated as fiscal news to identify unanticipated fiscal policy shocks.

### Empirical strategy (summary)
- Estimation method: Linear panel Local Projections (LPs) following Jordà (2005).
  - LP specification: y_{i,t+h} = α_i + β^h η_{i,t} + Ψ^h X_{i,t} + Φ^h(L) y_{i,t} + ε_{i,t+h}, h = 0, 1, 2, ..., H.
  - Inference with country-clustered standard errors.
- Projection horizon: 30 days (1 month); confidence intervals at 90 percent.
- Main dependent variables (in log deviations from t−1): COVID-19 cases, NO2 emissions (proxy for economic activity), 5-year sovereign CDS spreads (proxy for fiscal space).
- Shock variables: physical index, physical and smart index, smart index.
- Controls (X_{i,t}) include:
  - COVID-19 deaths, vaccinations, vaccination policy.
  - Temperature, humidity, pm10.
  - CDS spreads at 1, 3, 5, 10 years (in logs).
  - Sovereign yields at 1 and 10 years.
  - When dependent variable is economic activity or fiscal space, include COVID-19 cases as a control.
- Robustness: VIX included in robustness checks (results do not vary when VIX is included).
- Identification concerns:
  - Mitigation 1: Interact physical index with a seasonality factor (ratio of non-seasonally adjusted economic indicator to its seasonally adjusted counterpart) to account for seasonal variation and voluntary social distancing.
  - Mitigation 2: Include lags of dependent variable and country-specific time trends.
- State-dependent LPs: use dummy indicator D_t to separate regimes (eq. 2) and study heterogeneity across regimes such as advanced/emerging economies, slow/fast public health response time, high/low public debt.

### Baseline empirical results (key findings)
- General trade-off:
  - More physical lockdowns → lower COVID-19 cases but larger negative impact on economic activity.
  - Impacts are strongest for physical lockdowns compared to physical and smart, and smart-only measures.
- Quantitative headline effects (from unitary shocks in indices; indices scaled 0–1):
  - COVID-19 cases:
    - A unitary shock in the "physical" and "physical and smart" indices is associated with a 90 percent drop in COVID-19 cases at peak (stronger under physical).
    - A unitary shock in the "smart" index is associated with a 20 percent drop in COVID-19 cases at peak.
  - Economic activity (NO2 emissions):
    - A unitary shock in the "physical" index is associated with a peak 90 percent drop in NO2 emissions after a few days.
    - A unitary shock in the "physical and smart" index yields a milder plunge in NO2 emissions compared to physical-only.
    - A unitary shock in the "smart" index yields no statistically significant drop in NO2 emissions.
  - Fiscal space (5-year CDS spreads):
    - CDS spreads drop by about 20 percent on average across the three types of containment measures; IRFs are consistently negative throughout the projection horizon.
    - This corresponds to approximately a 25 basis points drop in CDS spreads in the sample.
- Interpretation:
  - Smart measures (testing, contact tracing, facial coverings) soften the health-economy trade-off: they are associated with milder reductions in COVID-19 cases but preserve economic activity and deliver statistically comparable reductions in sovereign risk relative to physical lockdowns.
  - Smart measures can be empirically associated with a slight reduction in COVID-19 infection cases while being least disruptive to economic activity and improving sovereign risk.

*Source: wpiea2022012-print-pdf - 3.  DATA*

### Appendix B.

### Appendix B

### Baseline results
- Responses are reported in log-deviations from past values. Shaded green areas represent the 90 percent confidence intervals; black solid lines represent median IRFs.
- Baseline findings:
  - Containment measures limit infection cases but can reduce economic activity.
  - Physical lockdowns are effective in reducing infection cases but can be disruptive to the economy.
  - "Smart" measures (e.g., testing and contact tracing) can control infections to some degree while safeguarding economic activity.
  - Smart measures are statistically as effective as physical ones in improving sovereign risk (reducing 5-year CDS spreads).
  - Implication: smart measures can contain infections while avoiding economic disruptions and improving sovereign risk.

### State-dependencies
- Method: state-dependent local projection (eq. 2); state dummy D_t equals 1 for the named state and 0 otherwise.

- Advanced versus Emerging economies
  - AEs typically have lower perceived sovereign risk than EMs.
  - Tightening of the physical index implies a stronger drop in NO2 emissions under the AE state, peaking at -100 percent, potentially reflecting weaker compliance in EMs.
  - CDS spreads under the AE regime tend to drop less than under the EM regime; in the EM regime spreads decrease consistently throughout the projection horizon.
  - Quantitative note: "That is equal to a 30 percent reduction of CDS spreads on average across the projection horizon."
  - Interpretation: financial markets may reward EMs more for taking difficult policy actions and/or sovereign risk in AEs is affected more by fundamentals.

- Fast versus Slow public health response time (PHRT)
  - PHRT definition (baseline): number of days needed to reach maximum stringency after a major COVID-19 outbreak (100 COVID-19 cases). State dummy equals 1 when PHRT is above the median (slow) and 0 if below the median (fast).
  - Findings:
    - Tightening of the physical index is associated with an initial drop in NO2 emissions in both regimes; the shock dissipates faster under the fast PHRT regime.
    - CDS spreads are statistically reduced only under the fast PHRT regime.
  - Robustness: alternative PHRT definitions considered:
    - Days to reach maximum stringency after 250 COVID-19 cases per 100 thousands inhabitants outbreak.
    - Days to reach a 1 percent tightening in the stringency index after 100 cases outbreak.

- High versus Low public debt
  - Two high-debt state definitions:
    - (i) debt-to-GDP above 90 percent of GDP cutoff;
    - (ii) observations above the 75th percentile of the public debt sample distribution.
  - Findings:
    - 5-year CDS spreads decrease significantly throughout the projection horizon only when public debt is low.
    - Peak response about -20 percent around 20 days after the shock (overall sample).
    - Considering EMs only, the decrease in CDS spreads peaks to -50 percent at the end of the projection horizon when public debt is low.
    - Implication: containment measures can be more effective in reducing sovereign risk when initial fiscal conditions are stronger.
  - Additional note: Figure 24 (Appendix B) reports that the cut in NO2 emissions is more subdued under low public debt.

### Fiscal announcements
- Sample and construction:
  - Daily frequency fiscal policy announcements series for EU-19 countries between February 2020-June 2021; series taken in percentage of GDP to be comparable across countries.
- Linear local projection results (eq. 1) — effect of fiscal announcement (FPA) shock on CDS and NO2 on impact and after 4 weeks (Table2):
  - Columns: CDS on impact | CDS after 4 weeks | NO2 on impact | NO2 after 4 weeks
  - Reported estimates for FPA:
    - CDS on impact: 0.0378 (t-statistic (0.93))
    - CDS after 4 weeks: 0.223 ∗∗∗ (t-statistic (3.18))
    - NO2 on impact: -0.201 (t-statistic (-1.14))
    - NO2 after 4 weeks: 2.070 ∗∗ (t-statistic (2.48))
  - Significance codes: * p<0.1, ** p<0.05, *** p<0.01
- Interpretation:
  - On average, an increase in fiscal support (fiscal announcement shock) is associated with a statistically significant mild rise in CDS spreads 4 weeks after the shock.
  - The same shock yields a doubling in NO2 emissions 4 weeks after the impulse shock.
  - Implication: fiscal policy interventions can support the economy but at the cost of slightly worsening sovereign risk.

- State-dependencies: large versus low fiscal announcements
  - Estimating eq. 2 to gauge impact of containment ("physical" and "smart") measures on CDS spreads conditional on fiscal announcement size.
  - Findings:
    - Size of fiscal announcements did not matter for the transmission of physical index shocks on CDS spreads (median effects negative).
    - Fiscal space improves when a mix of "large" fiscal support and "smart" measures are in place.
  - Interpretation: large support packages may be expected to worsen public finances and therefore future default risks embedded in 5-year CDS spreads; however, when combined with smart containment measures that rely on testing and avoid physical lockdowns, the confidence boost and more positive outlook can reduce expected sovereign risk.

### Concluding remarks
- Data and methods:
  - Daily data from February 2020 to June 2021 for 44 advanced and emerging economies.
  - Proxies: OxCGRT indices for containment measures, NO2 emissions for economic activity, CDS spreads for fiscal space.
  - Econometric approach: local projection à la Jordà (2005) and specification building on Deb et al. (2020a).
- Key summary findings:
  - Smart containment measures are relatively optimal: they can contain infections, avoid disruptions to economic activity, and improve sovereign risk.
  - State-dependent evidence:
    - In EMs versus AEs, stricter physical containment measures tend to affect economic activity less and reduce CDS spreads more.
    - Fast versus slow PHRT: faster response helps economic activity normalize more quickly and improves sovereign risk profile.
    - Low versus high initial public debt: containment measures are associated with improved economic outlook and lower sovereign risk when public debt is low.
  - Sovereign risk improves under a combination of large fiscal support and smart measures.
- Future research:
  - Further exploit and expand the fiscal announcement dataset to gauge effects of fiscal news on fiscal space and economic activity during COVID-19.

*Sources: Authors’ calculations.*

### REFERENCES

### REFERENCES (wpiea2022012-print-pdf)

### Bibliographic sources cited
- Extensive list of academic and policy references related to:
  - Epidemiological modeling and SIR frameworks (e.g., Kermack and McKendrick (1927); Acemoglu et al. (2020); Garibaldi, Moen, and Pissarides (2020); Bognanni et al. (2020)).
  - Empirical studies of COVID-19 containment measures and economic impact (e.g., Hsiang et al. (2020); Chinazzi et al. (2020); Kraemer et al. (2020); Tian et al. (2020); Coibion, Gorodnichenko, and Weber (2020)).
  - Fiscal space, fiscal limits, and sovereign risk analysis (e.g., Bi (2012); Bi and Leeper (2013); Bi and Traum (2012, 2014); Ghosh et al. (2013); Caselli et al. (2018); Ostry et al. (2010, 2015); Collard, Habib, and Rochet (2015); Botev, Fournier, and Mourougane (2016)).
  - COVID-19-specific fiscal, health, and vaccine research and IMF working papers (multiple works by Deb, Furceri, Ostry, Tawk and coauthors: IMF Working Paper Nos. 20/158, 20/159, 20/234, 20/125, 21/247, 21/248, and IMF Working Papers 2021(262)).
  - Methodological and econometric references (e.g., Jordà (2005); Plagborg-Møller and Wolf (2021); Granger and Teräsvirta (1993); Teräsvirta (1994); Ramey (2011a, 2011b); Ramey and Zubairy (2018); Tenreyro and Thwaites (2016)).

- Additional referenced studies and datasets include works on:
  - Financial market expectations and sovereign bond risk during COVID-19 (e.g., Ettmeier, Kim, and Kriwoluzky (2020); Andries, Ongena, and Sprincean (2021); Augustin et al. (2021); Cevik and Ozturkkal (2020); Esteves and Sussman (2020)).
  - Databases and data resources (e.g., Kose et al. (2017) cross-country fiscal space database; IMF’s Fiscal Monitor database; Yale’s COVID-19 Financial Response Tracker (CFRT); IMF Policy Tracker).

### Notable numeric and document references preserved exactly as cited
- Time period for fiscal announcements dataset construction: "Feb2020 to end-June2021".
- Country grouping: "EU-19".
- IMF Working Paper numbers and references preserved exactly where provided: e.g., "IMF Working Paper No. 20/234", "IMF Working Papers No. 20/125", "IMF Working Paper No. 20/158", "IMF Working Paper No. 20/159", "IMF Working Paper No. 21/99", "IMF Working Paper No. 21/247", "IMF Working Paper No. 21/248", "IMF Working Papers 2021(262)".
- World Economic Outlook citation: "WEO, I. (10/2020). World economic outlook, october."
- Classical epidemiology reference: "Kermack, W. O. and A. G. McKendrick (1927). A contribution to the mathematical theory of epidemics. Proceedings of the royal society of london. Series A ... 115(772), 700–721."

---

### APPENDIX A. FISCAL ANNOUNCEMENTS DATASET

### Dataset construction and sources
- Starting point:
  - Yale’s COVID-19 Financial Response Tracker (CFRT).
  - Methodology follows Deb et al. (2021).
- Supplementary and cross-checking sources:
  - IMF Policy Tracker.
  - Other reports (used especially when CFRT-reported numbers did not match other sources).
- Country coverage and allocation:
  - Focus on "EU-19" countries.
  - EU-level fiscal measures distributed to each country by each country’s GDP shares when measures were implemented at the EU level.
- Time coverage:
  - Daily high-frequency dataset covering "Feb2020 to end-June2021".

### Classification and aggregation
- Policy-instrument categorization and macro-aggregation follow IMF’s Fiscal Monitor database of country fiscal measures in response to the COVID-19 pandemic.
- Macro-categories used:
  - Lifeline and demand (spending and revenue) support.
  - Above-the-line and below-the-line measures.
  - Health and non-health measures.

### Definitions (as used in the dataset)
- Lifeline measures:
  - Include liquidity injections, loans in general, umbrella guarantees, credit guarantees, government provisions of loans, equity injections, asset purchases, targeted loans (support to damaged business/worst hit business).
- Demand-support measures:
  - Include wage subsidies, targeted transfers, grants, unemployment benefits, wage supplements, support to families with children, deferrals of tax and social security contributions, tax relief, social security support, and grants to small and medium enterprises (SMEs).
- Above-the-line measures:
  - Include unemployment benefits, grants and transfers, tax cuts and relief measures, tax or social security contribution payment deferrals, payment forbearances and support to SMEs (grants). These can be further categorized by health and non-health measures.
- Below-the-line measures:
  - Include any form of loans, capital injections, asset purchases and guarantees, government guarantees to banks, firms, households, and equity injections to firms.

### Measurement and interpretation
- Fiscal announcements are reported as the size of the shock in "percent of national GDP".
- The series of fiscal announcements are interpreted as unanticipated fiscal news/shocks in the same fashion as in Ramey (2011b,a).
- Figures referenced for Austria:
  - "Figure 8" reports raw fiscal announcements for Austria as reported in the CFRT.
  - "Figure 9" shows the expanded dataset after cross-checking and granular categorization.
  - "Figure 10" reports (static average) divisions across categories of fiscal announcements.
  - "Figure 11" shows the set of policy instruments under the demand support category.

- IMF’s Fiscal Monitor database access reference preserved exactly:
  - "IMF’s    Fiscal    Monitor    database    can    be    accessed    at    https://www.imf.org/en/Topics/imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19"

---

### APPENDIX B. ADDITIONAL RESULTS (FIGURES AND TABLES)

### Figures and their captions (overview of empirical outputs)
- Figures report impulse response functions (IRFs) and dynamic responses from linear local projections (eq. 1) for variables including:
  - New daily deaths, hospitalized patients, ICU patients (log-deviations).
    - "FIGURE12. IRFs of Covid-19-related deaths, hospitalized patients and ICU patients - Physical, physical and smart, smart indices."
    - Responses reported in log-deviations; shaded green areas denote "90% confidence intervals"; black solid lines represent median IRFs.
  - NO2 emissions (economic activity indicator) and 5-year CDS spreads (fiscal space indicator) using sub-indices of the stringency index (physical closures).
    - "FIGURE13. IRFs of NO2 emissions and CDS spreads - Stingency index: sub-indices."
    - Responses reported in log-deviations; shaded green areas denote "90% confidence intervals"; black solid lines represent median IRFs.
  - Comparisons across country groups:
    - Advanced vs emerging economies responses to "Physical and smart index" and "Smart index" shocks.
      - "FIGURE14. Advanced/emerging economies - Physical and smart index."
      - "FIGURE15. Advanced/emerging economies - Smart index."
    - Confidence intervals for advanced economies shaded in blue; for emerging economies shaded in red; median IRFs as black solid lines; responses in log-deviations.
  - Fast vs slow public health response time states:
    - Multiple figures for "Physical and smart index", "Smart index", and alternative definitions.
      - "FIGURE16" through "FIGURE23" present variations (including first and second alternative definitions) for physical, physical and smart, and smart indices.
    - Plots compare dynamic responses under "slow public health response" (confidence intervals shaded in blue) and "fast public health response" (confidence intervals shaded in red).
    - Median IRFs are black solid lines; responses reported in log-deviations.
  - High/low public debt states and NO2 emissions IRFs to physical/stringency shocks:
    - "FIGURE24. High/low public debt - IRFs of NO2 emissions - Physical index."
    - High public debt thresholds mentioned expressly: "90% threshold" and "75 pct threshold".
    - Also includes "High public debt (EMEs only − median)" and corresponding low public debt comparisons.
    - Confidence interval color conventions and median IRFs as above; responses in log-deviations.

### Presentation conventions used across figures
- Horizontal axis typically labeled "Days" with horizons shown (e.g., "05 10 15 20 25 30").
- Vertical axes labeled as "Log-difference" or "Log-deviations".
- Shaded areas denote "90% confidence intervals" unless otherwise specified.
- Black solid lines represent median IRFs in all figures.
- Figures are numbered sequentially from "FIGURE8" through "FIGURE24" (with earlier figures "FIGURE8" and "FIGURE9" associated with the Fiscal Announcements dataset for Austria and "FIGURE10" and "FIGURE11" describing category breakdowns).

---

*This content unit corresponds to the REFERENCES section and Appendices A and B of the source PDF (wpiea2022012-print-pdf).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022012-print-pdf.pdf_