## wpiea2024120-print-pdf

## Source details

**Canonical URL:** [wpiea2024120-print-pdf](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024120-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024120-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024120-print-pdf.pdf.json)

---

### Data
- Data sources and aggregation:
  - Green bond issuance data: Bloomberg L.P.
  - Country climate vulnerability and climate policy indices: ND-GAIN; Germanwatch.
  - Issuance data aggregated into quarterly intervals by jurisdiction.
  - Aggregated sample: 2,709 corporate green bonds issuance by 63 jurisdictions, including time frame from 2012 Q1 to 2022 Q3.
  - Key variable: date of the first issuance of a sovereign green bond. Dates and samples for various analyses are provided in Table A.1 in the appendix (as per source).
- Key issuance analysis variables (quarterly frequency):
  - Dependent variables:
    - Number of issuance: Number of corporate green bond issues. Source: Bloomberg L.P.
    - Share of green bond issuance: Size of corporate green bonds issuance relative to total bond issuance in percentage points – in US dollars using the FX rate at issuance where necessary. Sources: Bloomberg L.P.; Dealogic.
    - Share of verified green bonds: Number of corporate green bond verified by ESG assurance providers relative to total number of corporate green bonds in percentage points. Source: Bloomberg L.P.
  - Explanatory variables:
    - Sovereign debut: Dummy equal to one in or after the first sovereign green bond has been issued. Author's calculation.
    - Log(Cumulative green bond issuance): Log of cumulative amount of total green bond issuance – in log US dollars. Source: Bloomberg L.P.
  - Climate variables:
    - Vulnerability score: Values from 0 to 1. Source: ND-GAIN.
    - Climate change performance index (CCPI): Values from 0 to 100. Source: Germanwatch.
- Liquidity and yield data (daily frequency) — construction and sample:
  - Daily bid-ask spread data for corporate green bonds and pairing of corporate green bond yields with equivalent-maturity government conventional bonds.
  - Government bond yield data sourced from Bloomberg L.P.
  - Daily government yield curve constructed via linear interpolation or extrapolation subject to rules described in source.
  - Yield spread = corporate green bond yield minus interpolated government conventional bond yield.
  - Each sovereign green bond debut treated as an independent event; corporate green bond liquidity and yield data anchored in a window spanning from six weeks before to six weeks after the sovereign green bond issue date.
  - New variable: sovereign greenium — sovereign conventional bond yield minus sovereign green bond yield at the issue date.
  - Sample for liquidity and yield analysis: final dataset comprises daily green bond liquidity and yield data involving 14 jurisdictions with 851 green bond issues (both debut and subsequent).
- Key liquidity and yield dependent variables:
  - Bid ask spread: Spread between corporate green bond bid yield and ask yield in percentage points. Source: Bloomberg L.P.
  - Yield spread: Corporate green bond yield minus government conventional bond yield in percentage points. Source: Bloomberg L.P.

### Stylised facts and market context
- Sustainable bond market outstanding amounts:
  - Increased from $246 billion outstanding to around $4.3 trillion as of Q1 2024 – a more than 17-fold increase over a period of 7 years.
  - Early 2021 record quarterly issuance: number = 848; amount = $282 billion.
- Market composition shifts:
  - Shift away from corporates towards the official sector including sovereigns since the pandemic.
  - Sovereigns and other official sector issuers now account for about 45% of the amount outstanding and more than half of the issuance volume since the pandemic.
- Sovereign issuance characteristics:
  - In all countries that have issued sovereign green bonds, the average issuance size is more than 10 times larger than the average historical issuance size of corporate green bonds in the same market.
  - In the sample, 11 out of 14 green sovereign bond debuts were in local currency.
  - The average size of sovereign green bonds is 17x larger than the average size of corporate green bonds.
- Maturity structure:
  - Overwhelming majority of sustainable bonds issued within maturities of 3-7 years and 7-15 years, with a roughly even split between the two.
  - Corporate sustainable bonds: 49% have a maturity of less than seven years.
  - Sovereign issuers: nearly 60% of their issuances are longer than 15 years.
- External reviews, frameworks, and impact reporting:
  - Over 85% of green bonds (both sovereign and corporate) have external reviews.
  - Sovereign green bond issuance: almost all sovereign issuers solicit at least one external review provider, often multiple providers.
  - Over 70% of green bonds come with an issuance framework.
  - Only around 10% of green bonds come with an impact assessment.
  - Commitment to impact and allocation reporting remains limited among private issuers.

### Empirical strategy
- Hypotheses:
  - Hypothesis 1: Sovereign debut has a size effect on corporate green bond issuance (number and share of amount that is green), controlling for trends.
  - Hypothesis 2: Sovereign debut has a reporting and verification quality effect on corporate green bond issuance.
  - Hypothesis 3: Sovereign debut improves pricing and liquidity in the corporate green bond market.
- Approach:
  - Panel regressions with country fixed effects; event-study-like strategy using sovereign debut as shock.
  - Quarterly regressions for hypotheses 1 and 2; daily event-study regressions (−6 weeks to +6 weeks) for hypothesis 3.
  - Interactions of sovereign debut with CCPI and ND-GAIN vulnerability scores to assess conditional effects.

### Size effect — key regression findings (impact on number and amount of corporate green bonds)
- Number of corporate green bond issues (share of total amount that is green) — Whole sample (Observations: 2709):
  - Sovereign debut coefficients:
    - (1) 2.899 *** (0.000)
    - (2) 1.707 *** (0.000)
    - (3) 0.777 *** (0.006)
    - (4) 0.801 *** (0.005)
  - Cumulative log green bond issuance (t-1):
    - (1) 0.0845 *** (0.000)
    - (2) 0.0139 (0.207)
    - (3) 0.00007 (0.995)
  - R-squared variants: r2_w 0.0458, 0.0871, 0.143, 0.144.
- Number of issues — Advanced economies subsample (Observations: 1161):
  - Sovereign debut coefficients:
    - (5) 4.904 *** (0.000)
    - (6) 3.772 *** (0.000)
    - (7) 1.796 *** (0.000)
    - (8) 1.736 *** (0.000)
  - Cumulative log green bond issuance (t-1):
    - (5) 0.0959 *** (0.000)
    - (7) -0.0480 ** (0.024)
- Amount of corporate green bond issuance (share of issuance amounts) — Whole sample (Observations: 2709):
  - Sovereign debut coefficients:
    - (1) 4.164 *** (0.000)
    - (2) 2.930 *** (0.001)
    - (3) 1.002 (0.240)
    - (4) 1.072 (0.264)
  - Cumulative log green bond issuance (t-1):
    - (1) 0.132 *** (0.000)
    - (2) 0.0226 (0.464)
- Amount — Advanced economies subsample (Observations: 1161):
  - Sovereign debut coefficients:
    - (5) 5.315 *** (0.000)
    - (6) 4.028 *** (0.000)
    - (7) 1.142 (0.196)
    - (8) 0.858 (0.394)
  - Cumulative log green bond issuance (t-1):
    - (5) 0.168 *** (0.000)
    - (7) 0.0910 ** (0.041)
- Interpretation:
  - β1 (sovereign debut) is positive and statistically significant for both full and advanced-economy samples; magnitude declines when adding controls/time/country dummies.
  - Significance for issuance amount weakens with time dummies, consistent with strong global market expansion during sample period.

### Conditional effects: climate policy and vulnerability
- Marginal effects:
  - The sovereign debut effect on the number of corporate green bonds issued increases with CCPI (stronger climate policy → larger impact).
  - Corporates more likely to issue more green bonds after sovereign debut in countries with lower climate vulnerability (lower ND-GAIN score).
  - For issuance amount, marginal effect does not vary statistically significantly with climate conditions (Graph A1).

### Reporting and verification quality effect
- Main finding:
  - Larger share of corporate green bonds issued after sovereign debut used green verification (second party opinion or green bond certification).
- Whole sample (Observations: 2709):
  - Sovereign debut coefficients:
    - (1) 34.44 *** (0.000)
    - (2) 35.01 *** (0.000)
    - (3) 4.014 (0.123)
    - (4) 3.623 (0.171)
  - Constant (1): 17.94 *** (0.000)
- Advanced economies subsample (Observations: 1161):
  - Sovereign debut coefficients:
    - (5) 42.82 *** (0.000)
    - (6) 43.26 *** (0.000)
    - (7) -1.974 (0.632)
    - (8) -2.949 (0.481)
- Interpretation:
  - Increase in green verification after sovereign debut is significant in specifications without time dummies; result weakens with time dummies.
- Robustness:
  - Alternative definitions of green verification (second-party opinions, green certifications, audits) and alternative data sources (Bloomberg; Climate Bond Initiative) produce similar results.

### Liquidity and pricing effect — methodology
- Event-study regression using daily bid-ask spreads and yield spreads for corporate green bonds in a −6 week to +6 week window around sovereign issuance.
- Regressions include month-year time fixed effects, country fixed effects, and corporate issuer fixed effects in various specifications.
- Yield spread defined as corporate green bond yield minus government conventional bond yield.
- Sovereign greenium included in some specifications (sovereign conventional bond yield minus sovereign green bond yield at issuance).

### Liquidity (bid-ask spread) — empirical findings (Obs 31255)
- Sovereign debut coefficients (selected, Table 6):
  - (1) -0.00902 *** (0.000)
  - (2) -0.00198 *** (0.000)
  - (3) -0.00756 *** (0.000)
  - (4) -0.00290 *** (0.004)
  - (5) 0.000966 (0.345)
  - (6) -0.00198 *** (0.000)
- Interpretation:
  - In five out of eight specifications, sovereign debut dummy has statistically significant negative coefficient (reduced bid-ask spreads).
  - Largest estimated coefficients consistent with a reduction rounded up to 1 basis point.
- Fit statistics:
  - r2 values: 0.00227, 0.741, 0.113, 0.119, 0.746, 0.741, 0.124, 0.746.

### Pricing (yield spread) — empirical findings (Obs 31255)
- Sovereign debut coefficients (selected, Table 7):
  - (1) -0.0904 *** (0.000)
  - (2) -0.0679 *** (0.000)
  - (3) 0.0204 (0.613)
  - (4) -0.0677 *** (0.006)
  - (5) -0.0232 * (0.078)
  - (6) -0.0679 *** (0.000)
- Interpretation:
  - Yield spreads reduced subsequent to sovereign debut in six out of eight specifications.
  - Effects in significant specifications range between 2 to 9 basis points.
- Fit statistics:
  - r2 values: 0.000422, 0.922, 0.0371, 0.0409, 0.923, 0.922, 0.0416, 0.923.

### Addressing endogeneity — sovereign greenium (Table 8)
- Sovereign greenium correlated with reductions in bid-ask and yield spreads.
- Bid-ask regressions (Obs 31255) — selected:
  - Sovereign debut: (1) -0.0090 *** (0.000); (2) -0.0018 *** (0.001); (3) -0.0071 *** (0.000).
  - Sovereign greenium at issue: (2) -0.00876 ** (0.031); (3) -0.0583 *** (0.005); (4) -0.0288 * (0.058).
- Yield spread regressions (Obs 31255) — selected:
  - Sovereign debut: (5) -0.0673 *** (0.007); (6) -0.0666 *** (0.000).
  - Sovereign greenium at issue: (5) -1.349 *** (0.000).
- Interpretation:
  - After controlling for sovereign greenium, sovereign debut coefficient remains statistically significant with expected sign in three out of four specifications for both bid-ask and yield regressions — suggesting timing endogeneity is unlikely to drive main result.

### Sensitivity: foreign currency sovereign green bonds (Tables A.2 and A.3)
- Concern: foreign currency sovereign issues may target foreign markets and weaken domestic effects.
- Bid-ask spread (Obs 31255) — debut and debut_foreign examples:
  - debut: (1) -0.00768 *** (0.000); (2) -0.00204 *** (0.001).
  - debut_foreign: (1) -0.0127 *** (0.000); (5) -0.00623 *** (0.003).
  - Conclusion: effect is either the same or even stronger for foreign currency bonds in some specifications.
- Yield spread (Obs 31255) — debut_foreign examples:
  - debut_foreign: (1) -0.567 *** (0.000); (2) -0.115 *** (0.000); (3) -0.420 *** (0.000); (5) -0.207 *** (0.000).

### Subsequent sovereign issues — opposite or null effects (Obs 41496)
- Liquidity (bid-ask spread, Table 9) — subsequent coefficients:
  - (1) 0.00477 *** (0.000); (2) 0.00589 *** (0.000); (3) 0.00421 ** (0.024); (6) 0.00669 *** (0.000).
  - Interpretation: liquidity is, if anything, diminished by subsequent sovereign green bond issues (higher bid-ask spreads), statistically significant in 6 out of 8 specifications.
- Pricing (yield spread, Table 10) — subsequent coefficients:
  - (1) 0.119 *** (0.000); (2) 0.105 *** (0.000); (3) 0.596 *** (0.000); (4) 0.146 *** (0.000); (8) 0.122 *** (0.002).
  - Interpretation: post-debut sovereign green bond issues result in statistically significantly higher corporate green bond yield spreads in every regression.

### Robustness for Emerging Market Economies (EME) subsample (Table A.4)
- Observations: 1505 in reported specifications.
- Selected sovereign debut coefficients (examples):
  - 0.896 *** (0.000); 0.243 (0.345); -0.0312 (0.900); 3.002 ** (0.017).
  - Verification/quality coefficients: 25.31 *** (0.000); 26.90 *** (0.000); 6.993 ** (0.026); 7.360 ** (0.024).
- Cumulative log green bond issuance (t-1) selected coefficients:
  - 0.0431 *** (0.000); 0.0210 ** (0.028); 0.0092 (0.366); 0.108 ** (0.012).

### Key conclusions and policy-relevant implications
- Three primary findings:
  1. Size: Sovereign debut associated with an increase in number and volume of corporate green bond issues. Catalytic effect is stronger where climate policy (CCPI) is stricter or where climate vulnerability is lower.
  2. Quality: Sovereign entrants promote best practice in green verification and reporting, inducing corporate issuers to follow suit.
  3. Liquidity and pricing: Sovereign debut increases liquidity and diminishes yield spreads in corporate green bond markets; the same impact is not observed for subsequent sovereign green bond issues.
- Policy implications:
  - Sovereign entry into sustainable bond markets can spur private sustainable bond market development and improve market liquidity and pricing, even if sovereigns are latecomers.
  - There is scope for sovereign issuers to further improve market transparency, consistent with NGFS (2022) recommendations.
  - Possible policy measures include supervisory schemes for green verification providers and consideration of standardising or making mandatory impact reporting.

*Source: wpiea2024120-print-pdf, sections 2.1 "Data", 3.1 "Size effect", 3.3 "Liquidity and pricing effect", and related subsections.*

### 2.1 Data

### 2.1 Data

### Data sources and aggregation
- Green bond issuance data: Bloomberg L.P.
- Country climate vulnerability and climate policy indices: ND-GAIN; Germanwatch.
- Issuance data aggregated into quarterly intervals by jurisdiction.
- Aggregated sample: 2,709 corporate green bonds issuance by 63 jurisdictions, including time frame from 2012 Q1 to 2022 Q3.
- Key variable: date of the first issuance of a sovereign green bond. Dates and samples for various analyses are provided in Table A.1 in the appendix (as per source).

### Key issuance analysis variables (quarterly frequency)
- Dependent variables:
  - Number of issuance: Number of corporate green bond issues. Source: Bloomberg L.P.
  - Share of green bond issuance: Size of corporate green bonds issuance relative to total bond issuance in percentage points – in US dollars using the FX rate at issuance where necessary. Sources: Bloomberg L.P.; Dealogic.
  - Share of verified green bonds: Number of corporate green bond verified by ESG assurance providers relative to total number of corporate green bonds in percentage points. Source: Bloomberg L.P.
- Explanatory variables — sovereign debut and market trends:
  - Sovereign debut: Dummy equal to one in or after the first sovereign green bond has been issued. Author's calculation.
  - Log(Cumulative green bond issuance): Log of cumulative amount of total green bond issuance – in log US dollars. Source: Bloomberg L.P.
- Explanatory variables — country climate vulnerability and climate policy:
  - Vulnerability score: Score measures a country's exposure, sensitivity, and ability to adapt to the negative impact of climate change. Lower scores are better. Values from 0 to 1. Source: ND-GAIN.
  - Climate change performance index: Index evaluates a country's climate protection policy performance. Higher scores are better. Values from 0 to 100. Source: Germanwatch.

### Liquidity and yield data (daily frequency) — construction and sample
- Daily bid-ask spread data for corporate green bonds and pairing of corporate green bond yields with equivalent-maturity government conventional bonds.
- Government bond yield data sourced from Bloomberg L.P.
- Daily government yield curve constructed via linear interpolation or extrapolation:
  - Linear extrapolation used when the minimum maturity within the government yield data exceeds 0.
  - No extrapolation beyond the maximum maturity observed in the government yield data for any jurisdiction.
  - When remaining maturity of a corporate bond exceeds the maximum maturity in the government yield curve, government yields corresponding to the maximum maturity are applied.
- Yield spread calculated as corporate green bond yield minus interpolated government conventional bond yield in a given jurisdiction.
- Each sovereign green bond debut treated as an independent event; corporate green bond liquidity and yield data anchored in a window spanning from six weeks before to six weeks after the sovereign green bond issue date.
- New variable: sovereign greenium — defined as sovereign conventional bond yield minus sovereign green bond yield at the issue date.
- Sample for liquidity and yield analysis: final dataset comprises daily green bond liquidity and yield data involving 14 jurisdictions with 851 green bond issues (both debut and subsequent).
- Key liquidity and yield variables (Table 2):
  - Dependent variables:
    - Bid ask spread: Spread between corporate green bond bid yield and ask yield in percentage points. Source: Bloomberg L.P.
    - Yield spread: Corporate green bond yield minus government conventional bond yield in percentage points. Source: Bloomberg L.P.
  - Explanatory variables:
    - Sovereign debut: Dummy equal to one for periods occurring after the initial issuance of sovereign green bond. Source: Bloomberg L.P.
    - Subsequent issue: Dummy equal to one for periods occurring after the issuance of each sovereign green bond, excluding the debut issuance. Source: Bloomberg L.P.
    - Greenium at issue: Sovereign conventional bond yield minus sovereign green bond yield at issue date in percentage points. Source: Bloomberg L.P.

### Stylised facts and market context
- Sustainable bond market outstanding amounts:
  - Increased from $246 billion outstanding to around $4.3 trillion as of Q1 2024 – a more than 17-fold increase over a period of 7 years.
  - Early 2021 record quarterly issuance: number = 848; amount = $282 billion.
- Market composition shifts:
  - Shift away from corporates towards the official sector including sovereigns since the pandemic.
  - Sovereigns and other official sector issuers now account for about 45% of the amount outstanding and more than half of the issuance volume since the pandemic.
- Sovereign issuance characteristics:
  - In all countries that have issued sovereign green bonds, the average issuance size is more than 10 times larger than the average historical issuance size of corporate green bonds in the same market (Graph 3, left-hand panel).
  - In the sample, 11 out of 14 green sovereign bond debuts were in local currency.
  - The average size of sovereign green bonds is 17x larger than the average size of corporate green bonds.
- Maturity structure:
  - Overwhelming majority of sustainable bonds issued within maturities of 3-7 years and 7-15 years, with a roughly even split between the two.
  - Corporate sustainable bonds: 49% have a maturity of less than seven years.
  - Sovereign issuers: nearly 60% of their issuances are longer than 15 years.
- External reviews, frameworks, and impact reporting:
  - Over 85% of green bonds (both sovereign and corporate) have external reviews.
  - Sovereign green bond issuance: almost all sovereign issuers solicit at least one external review provider, often multiple providers.
  - Over 70% of green bonds come with an issuance framework.
  - Only around 10% of green bonds come with an impact assessment.
  - Commitment to impact and allocation reporting remains limited among private issuers.

### Empirical strategy
- Hypothesis framing:
  - Hypothesis 1: The sovereign debut has a size effect on corporate green bond issuance — test whether sovereign debut increases the number and the share of the amount of corporate bonds issued that is green, after controlling for existing trends.
  - Hypothesis 2: The sovereign debut has a reporting and verification quality effect on corporate green bond issuance — by setting a benchmark for stringent green reporting and verification mechanisms.
  - Hypothesis 3: The sovereign debut improves the pricing and liquidity in the corporate green bond market.
- Primary empirical approach:
  - Panel regressions with country fixed effects; strategy akin to an event study with the sovereign debut used as a shock.
  - General specification for hypotheses 1 and 2 (quarterly frequency):
    - Dependent variable Corp_c,t can be number of green corporate issues or share of issuance amount that is green (hypothesis 1), or the share of corporate green bonds having certain verification features (hypothesis 2).
    - Explanatory variables include a sovereign debut dummy D(SGB issuance_c,t≥j), MarketTrend_c,t−1 (time dummy or cumulative log amount of past green bond issuance), and country fixed effects C.
  - Interaction tests:
    - Interact sovereign debut with country climate characteristics to test whether the effect depends on climate-related risks or the strength of climate policies.
    - Climate measures used: Climate Change Performance Index (CCPI) by Germanwatch; a sub-component of ND-GAIN for climate vulnerability.
  - Tests of hypothesis 3:
    - Event study methodology based on Dittmar and Yuan (2008).
    - Use daily bid-ask spreads and yields from corporate bonds in a -6 week to +6-week window as dependent variables.
    - Independent variables include the sovereign debut dummy and fixed time effects for each month-year combination.
    - Frequency for these regressions is daily.

_Italic: Source — wpiea2024120-print-pdf, section 2.1 "Data" and related subsections._

### 3.1 Size effect: Impact of sovereign debut on corporate green bond

### 3.1 Size effect: Impact of sovereign debut on corporate green bond issuance

### Summary finding
- Simple summary statistics (Graph 5) indicate a very large increase in both the number and the issuance amount of corporate green bonds issued after the sovereign debut.
- The increase is much larger in emerging market economies, but from very low levels compared with advanced economies.

### Regression evidence — number of corporate green bond issues (Table 3)
- Dependent variable: the share of the total amount of corporate issuance in the jurisdiction that is green.
- Whole sample (columns (1)–(4)):
  - Sovereign debut coefficients:
    - (1) 2.899 *** (0.000)
    - (2) 1.707 *** (0.000)
    - (3) 0.777 *** (0.006)
    - (4) 0.801 *** (0.005)
  - Cumulative log green bond issuance (t-1):
    - (1) 0.0845 *** (0.000)
    - (2) 0.0139 (0.207)
    - (3) 0.00007 (0.995)
  - Constant:
    - (1) 1.074 *** (0.001)
    - (2) 0.279 *** (0.006)
    - (3) -2.43e-14 (1.000)
    - (4) 1.04e-14 (1.000)
  - Observations: 2709
  - Country dummy: No for (1),(3); Yes for (2),(4)
  - Time dummy: No for (1),(2); Yes for (3),(4)
  - R-squared: r2_w 0.0458, 0.0871, 0.143, 0.144; r2_o 0.0293, 0.126, 0.108, 0.0961; r2_b 0.00632, 0.252, 0.121, 0.00660
- Advanced economies subsample (columns (5)–(8)):
  - Sovereign debut coefficients:
    - (5) 4.904 *** (0.000)
    - (6) 3.772 *** (0.000)
    - (7) 1.796 *** (0.000)
    - (8) 1.736 *** (0.000)
  - Cumulative log green bond issuance (t-1):
    - (5) 0.0959 *** (0.000)
    - (6) -0.0306 (0.143)
    - (7) -0.0480 ** (0.024)
  - Constant:
    - (5) 1.622 *** (0.007)
    - (6) 0.401 * (0.056)
    - (7) -1.16e-13 (1.000)
    - (8) 6.66e-15 (1.000)
  - Observations: 1161
  - Country dummy: No for (5),(7); Yes for (6),(8)
  - Time dummy: No for (5),(6); Yes for (7),(8)
  - R-squared: r2_w 0.102, 0.142, 0.244, 0.245; r2_o 0.0825, 0.163, 0.152, 0.140; r2_b 0.0470, 0.237, 0.00944, 0.0658
- Interpretation:
  - The coefficient β1 associated with sovereign debut is positive and statistically significant for both the full sample and the advanced-economy subsample.
  - The size of the coefficient diminishes when additional control variables, or time or country dummies are included.
  - Regression results for the emerging market economies subsample are reported in Table A.4 in the Appendix (not reproduced here).

### Regression evidence — amount of corporate green bond issuance (Table 4)
- Dependent variable: share of issuance amounts of corporate green bonds.
- Whole sample (columns (1)–(4)):
  - Sovereign debut coefficients:
    - (1) 4.164 *** (0.000)
    - (2) 2.930 *** (0.001)
    - (3) 1.002 (0.240)
    - (4) 1.072 (0.264)
  - Cumulative log green bond issuance (t-1):
    - (1) 0.132 *** (0.000)
    - (2) 0.0226 (0.464)
    - (3) 0.00454 (0.905)
  - Constant:
    - (1) 1.441 *** (0.000)
    - (2) 0.135 (0.686)
    - (3) -4.48e-13 (1.000)
    - (4) 8.44e-15 (1.000)
  - Observations: 2709
  - Country dummy: No for (1),(3); Yes for (2),(4)
  - Time dummy: No for (1),(2); Yes for (3),(4)
  - R-squared: r2_w 0.0122, 0.0220, 0.0560, 0.0561; r2_o 0.00984, 0.0200, 0.0544, 0.0541; r2_b 0.00547, 0.0238, 0.0194, 0.00861
- Advanced economies subsample (columns (5)–(8)):
  - Sovereign debut coefficients:
    - (5) 5.315 *** (0.000)
    - (6) 4.028 *** (0.000)
    - (7) 1.142 (0.196)
    - (8) 0.858 (0.394)
  - Cumulative log green bond issuance (t-1):
    - (5) 0.168 *** (0.000)
    - (6) 0.0221 (0.535)
    - (7) 0.0910 ** (0.041)
  - Constant:
    - (5) 1.441 *** (0.000)
    - (6) -0.774 * (0.077)
    - (7) -6.80e-14 (1.000)
    - (8) -1.33e-15 (1.000)
  - Observations: 1161
  - Country dummy: No for (5),(7); Yes for (6),(8)
  - Time dummy: No for (5),(6); Yes for (7),(8)
  - R-squared: r2_w 0.0382, 0.0689, 0.171, 0.173; r2_o 0.0348, 0.0495, 0.165, 0.160; r2_b 0.0518, 0.00511, 0.00234, 0.0350
- Interpretation:
  - Results for issuance amount are broadly in line with those for number of issues.
  - Significance of β1 is weaker when time dummies are included, possibly because the corporate green bond market expanded rapidly globally during the sample period, making additional country-level effects harder to detect.

### Conditional effects: climate policy and climate vulnerability
- The marginal effect of sovereign debut on the number of corporate green bonds issued strengthens with the strength of countries’ climate policy performance (measured by CCPI).
  - The stronger a country’s climate policy, the larger the impact of sovereign debut on the number of corporate green bonds issued.
  - This is consistent with a positive signalling effect of sovereign green bond issuance that is amplified when governments have stronger climate policies.
- Corporates are more likely to issue more green bonds after the sovereign debut in countries that are less exposed to climate risks (lower climate vulnerability).
- Graph A1 (appendix) shows conditional marginal effects for issuance amount as dependent variable:
  - The marginal effect on issuance size does not seem to vary statistically significantly with countries’ climate conditions.

### Reporting and verification quality effect (Table 5)
- Focus: impact of sovereign debut on the use of green verification by corporate green bond issuers.
- Main result: A larger share of corporate green bonds issued after the sovereign debut have used green verification (second party opinion or green bond certification).
- Whole sample (columns (1)–(4)):
  - Sovereign debut coefficients:
    - (1) 34.44 *** (0.000)
    - (2) 35.01 *** (0.000)
    - (3) 4.014 (0.123)
    - (4) 3.623 (0.171)
  - Constant:
    - (1) 17.94 *** (0.000)
    - (2) 17.89 *** (0.000)
    - (3) -3.87e-12 (1.000)
    - (4) 1.39e-13 (1.000)
  - Observations: 2709
  - Country dummy: No for (1),(3); Yes for (2),(4)
  - Time dummy: No for (1),(2); Yes for (3),(4)
  - R-squared: r2_w 0.0647, 0.0647, 0.266, 0.266; r2_o 0.0503, 0.0503, 0.208, 0.208; r2_b 0.0192, 0.0192, 0.0192, 0.0192
- Advanced economies subsample (columns (5)–(8)):
  - Sovereign debut coefficients:
    - (5) 42.82 *** (0.000)
    - (6) 43.26 *** (0.000)
    - (7) -1.974 (0.632)
    - (8) -2.949 (0.481)
  - Constant:
    - (5) 28.43 *** (0.000)
    - (6) 28.38 *** (0.000)
    - (7) -3.19e-12 (1.000)
    - (8) 1.63e-13 (1.000)
  - Observations: 1161
  - Country dummy: No for (5),(7); Yes for (6),(8)
  - Time dummy: No for (5),(6); Yes for (7),(8)
  - R-squared: r2_w 0.0827, 0.0827, 0.398, 0.398; r2_o 0.0763, 0.0763, 0.323, 0.322; r2_b 0.0578, 0.0578, 0.0578, 0.0578
- Interpretation:
  - The increase in green verification usage after sovereign debut is statistically significant in specifications without time dummies.
  - The result is weakened statistically once time dummies are controlled for.

### Robustness checks
- Alternative definitions of green verification were tested: second-party opinions, green certifications, audits, and others.
- Data sources used: Bloomberg and Climate Bond Initiative.
- The results remain similar across these alternative definitions and data sources.

*Source: wpiea2024120-print-pdf - 3.1 Size effect: Impact of sovereign debut on corporate green bond*

### 3.3 Liquidity and pricing effect: Impact of sovereign debut on the

### 3.3 Liquidity and pricing effect: Impact of sovereign debut on the liquidity and yields in the corporate green bond market

### Theoretical background
- Existing literature documents that sovereign issuance in conventional markets increases liquidity and diminishes yield spreads in corporate bond markets (Yuan, 2005; Dittmar and Yuan, 2008), via benchmark government securities making corporate bond markets more complete and reducing adverse selection costs.
- Counterarguments include potential crowding-out of trading in existing securities (Subrahmanyam, 1991).
- Hypothesis 3: the benchmark effect found in conventional markets can be documented in green markets, particularly at the sovereign debut issue.

### Methodology
- Event study using daily bid-ask spreads and yield spreads for corporate green bonds of the same jurisdiction over a -6 week to +6 week window around sovereign issuance.
- Yield spread defined as corporate green bond yield minus government conventional bond yield.
- Regressions include month-year time fixed effects, country fixed effects, and corporate issuer fixed effects in various specifications.

### Empirical findings — Liquidity (bid-ask spread) (Table 6)
- Main patterns:
  - In five out of eight specifications, the sovereign debut dummy has a statistically significant negative coefficient, indicating reduced bid-ask spreads, ceteris paribus.
  - Largest estimated coefficients are consistent with a reduction in the bid-ask spread rounded up to 1 basis point.

- Selected coefficient estimates (Sovereign debut) and p-values from Table 6:
  - (1) -0.00902 *** (0.000)
  - (2) -0.00198 *** (0.000)
  - (3) -0.00756 *** (0.000)
  - (4) -0.00290 *** (0.004)
  - (5) 0.000966 (0.345)
  - (6) -0.00198 *** (0.000)
  - (7) 0.000736 (0.697)
  - (8) 0.000966 (0.345)

- Observations and model fit:
  - Obs 31255 in all specifications.
  - r2 values across specifications: 0.00227, 0.741, 0.113, 0.119, 0.746, 0.741, 0.124, 0.746.
  - r2_a (adjusted R-squared): 0.00224, 0.739, 0.112, 0.119, 0.744, 0.739, 0.122, 0.744.
  - r2_within: 0.00227, 0.000409, 0.000664, 0.000262, 0.000029, 0.000409, 0.000005, 0.000029.

### Empirical findings — Pricing (yield spread) (Table 7)
- Main patterns:
  - Yield spreads are reduced subsequent to the sovereign debut event in specifications with statistically significant coefficients (six out of eight specifications).
  - Effects in significant specifications range between 2 to 9 basis points.

- Selected coefficient estimates (Sovereign debut) and p-values from Table 7:
  - (1) -0.0904 *** (0.000)
  - (2) -0.0679 *** (0.000)
  - (3) 0.0204 (0.613)
  - (4) -0.0677 *** (0.006)
  - (5) -0.0232 * (0.078)
  - (6) -0.0679 *** (0.000)
  - (7) -0.0129 (0.779)
  - (8) -0.0232 * (0.078)

- Observations and model fit:
  - Obs 31255 in all specifications.
  - r2 values: 0.000422, 0.922, 0.0371, 0.0409, 0.923, 0.922, 0.0416, 0.923.
  - r2_a: 0.000390, 0.921, 0.0360, 0.0405, 0.922, 0.921, 0.0402, 0.922.
  - r2_within: 0.000422, 0.00294, 0.000008, 0.000242, 0.000100, 0.00294, 0.000002, 0.000100.

### Addressing potential endogeneity — Sovereign greenium (Table 8)
- Definition: sovereign greenium = sovereign conventional bond yield minus sovereign green bond yield at issuance (interpreted as a measure of relative demand for green bonds at issue).
- Key results when including sovereign greenium:
  - Sovereign greenium is associated with reductions in both the bid-ask spread and the yield spread.
  - Statistical significance: more common in bid-ask regressions (3 out of 4) vs. yield spread regressions (1 out of 4).
  - After controlling for sovereign greenium, the sovereign debut coefficient remains statistically significant with expected sign in three out of four specifications for both bid-ask and yield spread regressions — indicating that timing endogeneity is unlikely to drive the main result.

- Selected coefficients from Table 8:
  - Bid-ask spread regressions:
    - Sovereign debut: (1) -0.0090 *** (0.000); (2) -0.0018 *** (0.001); (3) -0.0071 *** (0.000); (4) 0.00117 (0.257).
    - Sovereign greenium at issue: (1) -0.00077 (0.883); (2) -0.00876 ** (0.031); (3) -0.0583 *** (0.005); (4) -0.0288 * (0.058).
    - Obs 31255; r2: 0.00227, 0.741, 0.114, 0.746; r2_a: 0.00221, 0.739, 0.112, 0.744; r2_within: 0.00227, 0.000559, 0.000912, 0.000145.
  - Yield spread regressions:
    - Sovereign debut: (5) -0.0673 *** (0.007); (6) -0.0666 *** (0.000); (7) 0.0227 (0.575); (8) -0.0222 * (0.092).
    - Sovereign greenium at issue: (5) -1.349 *** (0.000); (6) -0.0817 (0.115); (7) -0.265 (0.602); (8) -0.136 (0.484).
    - Obs 31255; r2: 0.00439, 0.922, 0.0371, 0.923; r2_a: 0.00433, 0.921, 0.0360, 0.922; r2_within: 0.00439, 0.00302, 0.000017, 0.000116.

### Sensitivity: foreign currency sovereign green bonds (Tables A.2 and A.3)
- Concern: foreign currency sovereign issues may target foreign markets, potentially weakening effects on domestic corporate issuers.
- Findings (Table A.2 — bid-ask spread):
  - debut coefficients: (1) -0.00768 *** (0.000); (2) -0.00204 *** (0.001); (3) -0.00745 *** (0.000); (4) -0.00267 ** (0.013); (5) 0.00175 * (0.098); (6) -0.00204 *** (0.001); (7) 0.00184 (0.345); (8) 0.00175 * (0.098).
  - debut_foreign coefficients: (1) -0.0127 *** (0.000); (2) 0.000566 (0.755); (3) -0.00173 (0.546); (4) -0.00224 (0.498); (5) -0.00623 *** (0.003); (6) 0.000566 (0.755); (7) -0.00875 ** (0.023); (8) -0.00623 *** (0.003).
  - Observations 31255 in all specifications.
  - Conclusion: effect is either the same or even stronger for foreign currency bonds in some specifications.
- Findings (Table A.3 — yield spread):
  - debut coefficients: (1) -0.0304 (0.235); (2) -0.0558 *** (0.000); (3) 0.0471 (0.245); (4) -0.0503 * (0.053); (5) 0.00290 (0.831); (6) -0.0558 *** (0.000); (7) 0.0179 (0.706); (8) 0.00290 (0.831).
  - debut_foreign coefficients: (1) -0.567 *** (0.000); (2) -0.115 *** (0.000); (3) -0.420 *** (0.000); (4) -0.165 ** (0.039); (5) -0.207 *** (0.000); (6) -0.115 *** (0.000); (7) -0.244 *** (0.010); (8) -0.207 *** (0.000).
  - Observations 31255 in all specifications.

### Subsequent sovereign issues: opposite or null effects (Tables 9 and 10)
- Liquidity (Table 9 — bid-ask spread for subsequent sovereign issuance):
  - Subsequent coefficients (1) 0.00477 *** (0.000); (2) 0.00589 *** (0.000); (3) 0.00421 ** (0.024); (4) 0.00591 *** (0.000); (5) 0.00374 ** (0.024); (6) 0.00669 *** (0.000); (7) 0.00146 (0.561); (8) 0.00126 (0.468).
  - Obs 41496 in all specifications.
  - Interpretation: liquidity is, if anything, diminished by subsequent sovereign green bond issues (higher bid-ask spreads), statistically significant in 6 out of 8 specifications.
- Pricing (Table 10 — yield spread for subsequent sovereign issuance):
  - Subsequent coefficients (1) 0.119 *** (0.000); (2) 0.105 *** (0.000); (3) 0.596 *** (0.000); (4) 0.146 *** (0.000); (5) 0.158 *** (0.000); (6) 0.116 *** (0.000); (7) 0.150 *** (0.005); (8) 0.122 *** (0.002).
  - Observations 41496 in all specifications.
  - Interpretation: post-debut sovereign green bond issues result in statistically significantly higher corporate green bond yield spreads in every regression.

### Robustness for Emerging Market (EME) subsample (Table A.4)
- Selected EME subsample results:
  - Sovereign debut coefficients across various hypotheses/specifications include:
    - 0.896 *** (0.000)
    - 0.243 (0.345)
    - -0.0312 (0.900)
    - 0.0782 (0.763)
    - 3.002 ** (0.017)
    - 1.829 (0.245)
    - 0.533 (0.702)
    - 1.095 (0.493)
    - 25.31 *** (0.000)
    - 26.90 *** (0.000)
    - 6.993 ** (0.026)
    - 7.360 ** (0.024)
  - Cumulative log green bond issuance (t-1) coefficients (selected): 0.0431 *** (0.000); 0.0210 ** (0.028); 0.0092 (0.366); 0.108 ** (0.012).
  - Observations 1505 in all specifications shown.
  - r2_w, r2_o, r2_b reported per specification; details preserved in table.

### Key conclusions and policy-relevant implications (from section 4 Conclusion)
- Three primary findings:
  1. Size: Sovereign debut associated with an increase in number and volume of corporate green bond issues. The catalytic effect is stronger the stricter a country’s climate policy or the less vulnerable the country is to climate risks.
  2. Quality: Sovereign entrants promote best practice in green verification and reporting, inducing corporate issuers to follow suit.
  3. Liquidity and pricing: Sovereign debut is a distinctive event that increases liquidity and diminishes yield spreads in corporate green bond markets; the same impact is not observed for subsequent sovereign green bond issues.
- Policy implications:
  - Sovereign entry into sustainable bond markets can spur private sustainable bond market development and improve market liquidity and pricing, even if sovereigns are latecomers.
  - There is scope for sovereign issuers to further improve market transparency, consistent with NGFS (2022) recommendations.
  - Possible policy measures include supervisory schemes for green verification providers and consideration of standardising or making mandatory impact reporting.

*Source: Sovereign Green Bonds: A Catalyst for Sustainable Debt Market Development? Working Paper No. WP/2024/120*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024120-print-pdf.pdf_