## Annex I : Description of Variables

## Source details

**Canonical URL:** [Annex I : Description of Variables](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020162-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020162-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020162-print-pdf.pdf.json)

---

### I. Introduction — Rationale and scope
- Study objective: investigate the effect on bond pricing of CACs and enhanced CACs embedded in international sovereign bond contracts using a comprehensive set of secondary-market bond yield spread data and some stylized primary-market yield observations.
- Sample period: May 1996–March 2020.
- Key innovation: systematic examination of the sovereign bond pricing impact of the inclusion of the ICMA’s 2014 model of enhanced CACs using comprehensive secondary-market sovereign bond yield spread data.
- Main high-level findings:
  - Inclusion of regular CACs is associated with lower costs most of the time.
  - Inclusion of enhanced CACs is associated with lower spreads since the introduction of this clause in August 2014.
- Policy relevance: provides evidence that market participants view CACs and enhanced CACs as facilitating orderly restructurings rather than increasing moral hazard.

### II. Literature review — Competing hypotheses and prior evidence
- Competing views:
  - CACs increase yields by making restructurings easier and reinforcing debtor moral hazard.
  - CACs decrease yields by facilitating orderly and efficient restructurings, reducing holdout problems, and speeding recovery and market access.
- Select empirical findings cited:
  - Becker and others (2003): no evidence CACs increased borrowing costs regardless of issuer rating.
  - Gugiatti and Richards (2003): no impact of CAC inclusion on yields for several emerging markets.
  - Bardozzetti and Dottori (2013): little impact for high- and low-rated issuers; reduce costs for mid-rated issuers (BB+ to B+).
  - Bradley and Gulati (2013): inclusion associated with lower borrowing cost, especially for financially weak issuers.
  - Stolper and Dougherty (2017): conclude no discernable CAC-related bond price effects.
  - Carletti and others (2016) on Venezuela (2010–16): bonds with no CACs requiring 100 percent vote were cheaper than bonds with 75 percent and 85 percent CACs; CAC inclusion positively associated with secondary-market yields during near-default situations (example: probability of default over 90 percent in June 2016).
- Empirical challenge: many nonlegal factors affect bond prices (geopolitical risks, negotiating capacity, likelihood of IMF bailout, liquidity, investor base structure, investor relations, credit ratings, haircut history).

### III. Overview and evolution of Collective Action Clauses (CACs)
- Historical and structural notes:
  - Prior to 2003 New York-law bonds generally did not include CACs; CACs traditionally included under English law.
  - Wide use of CACs began with Mexico in February 2003; inclusion became market practice for New York-law-governed bonds thereafter.
  - Typical required vote: 75 percent; actual required votes vary from 18.75 to 85 percent of outstanding bondholders (18.75 percent typically applied only if an initial quorum requirement is not satisfied).
  - Euro-area developments:
    - October 2010: standardized “double-limb” aggregation Euro CACs initiated for new euro area government bonds with maturity above one year, effective January 1, 2013.
    - Double-limb structure: (1) 75 percent across all series and (2) 66.67 percent in each series; series failing 66.67 percent are excluded.
    - Single-limb aggregation (ICMA enhanced CACs) prevents holdouts across series and is viewed as more effective than double-limb.
  - ICMA recommendations: August 2014 model enhanced CACs and model pari passu clause; IMF Executive Board endorsement in October 2014.
  - November 2018: Eurogroup announced broad support to amend ESM treaty to require single-limb CACs in all euro area issuances by 2022.
- ICMA model single-limb clauses key features:
  - (1) a “uniformly applicable” requirement in a single-limb voting procedure;
  - (2) a 75 percent aggregate voting requirement; and
  - (3) sub-aggregation.
- Market adoption and outstanding stock statistics (as of end-March 2020 unless otherwise noted):
  - Estimated outstanding foreign law-governed sovereign bonds: approximately $1.3 trillion.
  - Governing-law split: approximately 46 percent English law and approximately 52 percent New York law.
  - Inclusion of enhanced CACs: approximately 51 percent of the outstanding stock includes the ICMA’s enhanced CACs.
  - Old forms (two-limb aggregated or series-by-series CACs): 45 percent of the outstanding stock.
  - No CACs: 4 percent of the outstanding stock.
  - Characteristics of outstanding bonds without CACs:
    - About 44 percent of outstanding bonds without any CACs are below investment-grade.
    - Outstanding bonds without CACs would not mature until 2096.
    - 75 percent of outstanding bonds without CACs are under New York law.
  - Pari passu clause: about 50 percent of outstanding stock includes the ICMA’s strengthened pari passu clause.
- Data sources and estimation notes:
  - Share of outstanding stock calculated based on outstanding amount in US dollars as of end-March 2019 (data used in IMF’s Fourth Progress Report on Inclusion of Enhanced Contractual Provisions in International Sovereign Bond Contracts (2019)).
  - Data sources: Bloomberg, Dealogic, Perfect Information database, various countries’ authorities; excludes GDP warrants and China’s domestic issuances under Hong Kong SAR governing law.
- Figure 2 listed values (Total International Portfolio Investment Assets; units: trillions of US dollars; as of June 2019): 12.5, 4.8, 4.5, 3.6, 3.5, 3.4, 3.0, 2.4, 2.1, 1.8.

### IV. Data and sample
- Sample: 1,025 outstanding foreign law-governed sovereign bonds with secondary-market yields available at end of March 2020.
- Time coverage: monthly series from May 1996 to March 2020 (287 time-points).
- Countries: bonds from 116 countries for all regions (advanced economies and emerging markets); emerging markets broadly defined, including frontier markets and low-income countries.
- Bond selection rules:
  - Omitted bonds with remaining maturity of less than one year.
  - Excluded spreads over 1,000.
  - Included mostly conventional-type bonds; excluded complex coupon structures, convertibles, or variable rates.
  - Treated central bank bonds issued to finance the sovereign balance sheet as equivalent to government bonds issued by the ministry of finance.
  - Excluded state-owned enterprise bonds and government guaranteed bonds.
  - Included sukuk issued in international markets using their regular daily yields.
- CAC coverage in sample: 4 percent bonds with no CACs, 45 percent with regular CACs, and 51 percent with enhanced CACs.
- Primary pricing sources: BVAL and CBBT; fallback to Bloomberg generic BGN or others.
- Data sources: Bloomberg, IMF World Economic Outlook (WEO) database, Perfect Information database, Dealogic.
- Legal indexing: detailed indexing of legal clauses (various forms of CACs, enhanced CACs, pari passu and strengthened pari passu clauses) developed for this study.

### V. Methodology and variables
- Dependent variable: sovereign bond yield spread (basis points) over relevant currency benchmarks (10-year U.S. government treasury, German bund, Japanese government bond yields).
- Mid-yields to maturity: simple average of daily series, then converted to spreads versus corresponding benchmarks.
- Regression framework: panel regression model with bond-specific time-invariant effects.
  - Specification: Yi,t = α + β1 CAC + β2 CAC_ENHANCED + β3 Xi,t + β4 θi + εi,t
  - CAC variables: three mutually exclusive binary dummies — no CACs (1 or 0), regular CACs (1 or 0), enhanced CACs (1 or 0).
  - Interpretation example: a regular CAC coefficient of -40 implies the yield of bonds with CACs is 40 basis points lower than bonds without CACs.
- Independent variables (Xi,t) include:
  - Bond-specific: coupon, tranche volume in billions of US dollars, tenor, governing law (binary), SEC registration, stock exchange, issuing currency, origin maturity.
  - Time-variant macro/market: inflation (annual percent changes), sovereign five-year CDSs, changes in numeric credit ratings, remaining maturities (years), bid-ask spreads, VIX, exchange rates, debt-to-GDP ratios, EMBI spreads.
- Estimation details:
  - Monthly series used to reduce daily-yield noise.
  - Models run for whole sample and for subperiods; omitted variables when encountering near singular matrices.
  - Credit rating coding: numeric conversion where “lower the grade, the larger the numbers” (example: AAA coded 1, C coded 24).

### VI. Key empirical findings on CACs (secondary-market spreads)
- Overall effect:
  - The presence of CACs is associated with lower secondary-market sovereign bond yield spreads.
  - Relationship more pronounced for noninvestment-grade issuers.
- Regular CACs:
  - Most of the time, regular CACs are negatively associated with bond yield spreads, except during the European sovereign debt crisis period.
  - For noninvestment-grade issues, the negative association with spreads is more consistent.
  - For investment-grade issues, regular CACs have a positive coefficient (associated with higher borrowing costs) across some periods.
- Enhanced CACs (ICMA introduction in August 2014):
  - Bonds that included enhanced CACs exhibit negative coefficients with spreads and these results are consistently statistically significant.
  - Presence of enhanced CACs is associated with lower secondary-market yield spreads.
  - Findings suggest investors do not price bonds adversely based solely on inclusion of enhanced CACs.

### VII. Periodized empirical results (selected estimates and diagnostics)
- Study periods:
  - May 1996–January 2003 (Before New York law–governed sovereign bonds included CACs)
  - February 2003–May 2007 (Pre-global financial crisis)
  - June 2007–December 2009 (Global financial crisis)
  - January 2010–August 2014 (European sovereign debt crisis)
  - September 2014–March 2020 (After ICMA introduced enhanced CACs)
- Pre-February 2003:
  - CACs estimated to have contributed to lowering yield spreads by 1.3 percent (statistically significant) for investment-grade issuers; effect insignificant for noninvestment-grade issuers.
- February 2003–May 2007:
  - Spreads of bonds with CACs lower than those without CACs for the entire sample; effects statistically significant for both noninvestment-grade and investment-grade issuers.
  - Reported R-squared values: 0.42 (Total), 0.34 (Non-investment grade), 0.66 (Investment grade).
- June 2007–December 2009:
  - Inclusion of CACs appears to decrease bond yield spreads for noninvestment-grade countries by 402 basis points (statistically significant).
  - For investment-grade issuers, CACs are positively associated with spreads (statistically significant).
  - Reported R-squared values: 0.97 (Total), 0.89 (Non-investment grade), 0.90 (Investment grade).
- January 2010–August 2014:
  - Inclusion of CACs associated with slightly higher borrowing costs for the overall sample; for both investment-grade and noninvestment-grade issuers CACs associated with higher bond yield spreads (statistically significant).
  - Reported R-squared values: 0.87 (Total), 0.91 (Non-investment grade), 0.76 (Investment grade).
- September 2014–March 2020 (post-ICMA):
  - Enhanced CACs coefficient estimates (Table 7): -17.9 (Total) **, -129.8 (Non-investment grade) **, -16.6 * (Investment grade).
  - Regular CACs coefficient estimates (Table 7): -31.9 ** (Total), -85.9 ** (Non-investment grade), -34.5 ** (Investment grade).
  - R-squared values: 0.73 (Total), 0.45 (Non-investment grade), 0.64 (Investment grade).
  - Notes: **99% significant, * 95% significant.
- Venezuela pricing data for September 2014–March 2020 excluded due to highly volatile secondary-market trading activity.

### VIII. Determinants of secondary-market yield spreads (selected results)
- Credit ratings (time-variant):
  - Strong explanatory variable. More creditworthy bonds (numeric values 1–10) associated with lower yield spreads; less creditworthy bonds (numeric values over 10) associated with higher yield spreads.
- Remaining maturities:
  - Positively associated with spreads.
- Bond-specific characteristics:
  - Coupon and tranche volumes: positively associated with yield spreads; tranche volumes important for yields.
  - Issue currency: issuing in euros or US dollars associated with lower yield spreads compared to Japanese yen, British pound, Scandinavian currencies, Chinese renminbi.
  - Governing law: not consistently significant over full sample; English governing law associated with higher yield spreads only in the latest periods (after 2014) and statistically significant then.
- Liquidity (bid-ask spreads) and SEC registration:
  - Bid-ask spreads positively associated with yield spreads (greater liquidity — lower bid-ask spread — associated with lower yield spreads).
  - SEC registration shows a statistically strong negative relationship with yield spreads since 2010.
- Macroeconomic and market variables:
  - CPI annual growth rate: positively associated with yield spreads.
  - CDS spread and market VIX: positively associated with yield spreads.
  - Commodity prices: not statistically significant, mixed signs.
  - Monetary union membership: associated with higher secondary-market bond yield spreads during the European sovereign debt crisis.
- General conclusion: macroeconomic fundamentals and market variables play a major role in explaining government bond yield differentials.

### IX. Annex II — Alternative regression results (selected coefficients; period: 2014M9-2020M3)
- Purpose: limit sample to U.S. dollar-denominated sovereign bond issuances; eliminate potentially endogenous independent variables; exclude spreads over 1,000 basis points.
- Main inference:
  - Inclusion of single-limb CACs (ICMA enhanced CACs) is associated with lower borrowing costs for both noninvestment- and investment-grade issuers.
  - Inclusion of regular CACs is associated with even lower borrowing costs.
  - Noninvestment-grade issuers benefit more from including single-limb CACs than investment-grade issuers.
  - Including bonds with spreads over 1,000 basis points yields similar coefficient results but R-squared falls drastically to 22 percent.
- Selected panel regression coefficient estimates (Dependent variable: Spread over benchmark in bps; columns: Total | Non-investment grade | Investment grade):
  - Coupon: 27.7 ** | -5.9 | 11.8 **
  - Outstanding $: -52.4 | -111.2 | 19.9 **
  - Maturity: 3.4 ** | 4.5 ** | 2.6 **
  - NY law: -41.4 ** | -160.0 ** | -18.7 **
  - SEC: -15.4 ** | 42.7 ** | -42.0 **
  - EMBI: 0.5 ** | 0.4 ** | 0.3 **
  - VIX: 1.1 ** | 2.3 ** | 1.4 **
  - Exchange rate: 0.0 ** | 0.0 | 0.0 **
  - CPI(-6): 0.1 | -1.5 | -3.3
  - DEBT(-6): 0.3 * | 1.6 * | -1.6 *
  - Enhanced CACs: -29.6 * | -70.5 * | -35.4 **
  - Regular CACs: -59.9 ** | -113.7 ** | -28.9 **
  - C (constant): -676.0 | -869.2 | -515.9
  - R-squared: 0.66 | 0.48 | 0.69
- Notes: **99% significant, * 95% significant.

### X. Credit-rating numeric conversion (Annex II. Table 2)
- Composite credit ratings → Numeric value:
  - AAA → 1
  - AA+ → 2
  - AA → 3
  - AA- → 4
  - A+ → 5
  - A → 6
  - A- → 7
  - BBB+ → 8
  - BBB → 9
  - BBB- → 10
  - BB+ → 11
  - BB → 12
  - BB- → 13
  - B+ → 14
  - B → 15
  - B- → 16
  - CCC+ → 17
  - CCC → 18
  - CCC- → 19
  - CC+ → 20
  - CC → 21
  - CC- → 22
  - C+ → 23
  - C → 24
  - C- → 25
  - D → 26
  - Not rated → 27

### XI. Primary-market observations (selected cases)
- Case selection: Mexico (New York law, investment-grade), Romania (UK law, investment-grade), Indonesia (New York law, noninvestment-grade).
- Finding: Yields of bonds with enhanced CACs at issuance generally align with the sovereign yield curve of the issuer; no obvious pricing premium for enhanced CACs at issuance observed across these selected cases.

### XII. Caveats, limitations, and directions for further research
- Limitations noted:
  - Composition of investor base can change quickly and affect pricing.
  - Variations in CAC formulations (minimum voting requirements of 66.6 percent, 75 percent, 85 percent, and 100 percent; mandatory meeting requirements; aggregation; acceleration; reverse acceleration; collective representation) were not distinguished.
  - Other contractual provisions (pari passu/modified pari passu, cross-default clauses, engagement clauses, disenfranchised clauses, information covenant clauses) may influence pricing and were not separately analyzed.
  - Variations within rating groups and issuer structures (trustee vs fiscal agency) were not differentiated.
  - Treatment of restructured/exchanged bonds (different payment schedules, amortization) may affect pricing and warrants further study.

### XIII. Overall conclusion (as presented)
- Using an extensive secondary-market bond spread dataset, the study finds:
  - Regular CACs lowered the cost of borrowing for most considered historical periods, with effects more pronounced for noninvestment-grade issuers.
  - The bond-yield-spread/borrowing-cost-reducing effects are strongest in the first two periods (pre-2003 and 2003–2007) and lose statistical significance starting with the global financial crisis.
  - Enhanced (single-limb) CACs are associated with lower bond yield spreads.
- Interpretation: results consistent with secondary-market investors’ implied acknowledgement of CACs’ potential benefits to an efficient and orderly restructuring process, helping especially the pricing of noninvestment-grade issues.

*Source: wpiea2020162-print-pdf - Annex I : Description of Variables (excerpt provided).*

### Annex I : Description of Variables .....................................................................................

### Annex I : Description of Variables

### I. Introduction — Rationale and scope
- Study objective: investigate the effect on bond pricing of CACs and enhanced CACs embedded in international sovereign bond contracts using a comprehensive set of secondary-market bond yield spread data and some stylized primary-market yield observations.
- Sample period noted: May 1996–March 2020.
- Key innovation: systematic examination of the sovereign bond pricing impact of the inclusion of the ICMA’s 2014 model of enhanced CACs using comprehensive secondary-market sovereign bond yield spread data.
- Main high-level finding stated:
  - Inclusion of regular CACs is associated with lower costs most of the time.
  - Inclusion of enhanced CACs is associated with lower spreads since the introduction of this clause in August 2014.
- Policy relevance: sheds light on IMF Progress Reports’ remark that inclusion of enhanced CACs “doesn’t seem to have an observable pricing effect” (IMF 2019), providing evidence that market participants view CACs and enhanced CACs as facilitating orderly restructurings rather than increasing moral hazard.

### II. Literature review — Competing hypotheses and prior evidence
- Two broad opposing views on cost impact of CACs:
  - CACs increase yields: make restructurings easier, reinforce debtor moral hazard, investors demand higher yields (cited arguments).
  - CACs decrease yields: facilitate orderly and efficient restructurings, reduce holdout problems, speed recovery and market access, increasing expected long-term returns.
- Empirical literature summarized (select findings as presented):
  - Becker and others (2003): no evidence CACs increased borrowing costs regardless of issuer rating.
  - Gugiatti and Richards (2003): no impact of CAC inclusion on yields for several emerging markets.
  - Bardozzetti and Dottori (2013): little impact for high- and low-rated issuers; reduce costs for mid-rated issuers (BB+ to B+).
  - Bradley and Gulati (2013): inclusion associated with lower borrowing cost, especially for financially weak issuers.
  - Stolper and Dougherty (2017): argue similar lines to the IMF Progress Report—no discernable CAC-related bond price effects.
  - Carletti and others (2016) on Venezuela: during distress (2010–16), bonds with no CACs requiring 100 percent vote were cheaper than bonds with 75 percent and 85 percent CACs; CAC inclusion positively associated with secondary-market yields during near-default situations (example: probability of default over 90 percent in June 2016).
- Empirical challenge emphasized: many nonlegal factors affect bond prices (geopolitical risks, negotiating capacity, likelihood of IMF bailout, liquidity, investor base structure, investor relations, credit ratings, haircut history).

### III. Overview and evolution of Collective Action Clauses (CACs)
- Historical notes and forms:
  - Sovereign debt issuances prior to 2003 under New York law did not generally include CACs; CACs traditionally included under English law.
  - Wide use of CACs started with Mexico in February 2003; inclusion became market practice for New York-law-governed bonds since then.
  - Typical required vote: 75 percent, but “required votes” vary from 18.75 to 85 percent of outstanding bondholders.
    - Note: 18.75 percent vote typically applied only if an initial quorum requirement is not satisfied.
  - Euro area developments:
    - October 2010: standardized “double-limb” aggregation Euro CACs initiated for new euro area government bonds with maturity above one year, effective January 1, 2013.
    - Double-limb structure: requires support both (1) across all series being restructured (75 percent); and (2) in each series (66.67 percent). If a series does not meet 66.67 percent, it is excluded.
    - Single-limb aggregation (ICMA enhanced CACs) prevents holdouts across series and is viewed as more effective than double-limb for addressing collective action problems.
  - ICMA recommendations: August 2014 model enhanced CACs and model pari passu clause; IMF Executive Board endorsement in October 2014.
  - November 2018: Eurogroup announced broad support to amend ESM treaty to require single-limb CACs in all euro area issuances by 2022 (current ESM treaty requires double-limb CACs for issuances after January 2013).
- ICMA model single-limb clauses key features (as listed):
  - (1) a “uniformly applicable” requirement in a single-limb voting procedure;
  - (2) a 75 percent aggregate voting requirement; and
  - (3) sub-aggregation.
- Market adoption and outstanding stock statistics (as of end-March 2020 unless otherwise noted):
  - Estimated outstanding foreign law-governed sovereign bonds: approximately $1.3 trillion.
  - Governing-law split of that stock: approximately 46 percent governed by English law and approximately 52 percent by New York law.
  - Inclusion of enhanced CACs: approximately 51 percent of the outstanding stock includes the ICMA’s enhanced CACs.
  - Old forms (two-limb aggregated or series-by-series CACs): 45 percent of the outstanding stock.
  - No CACs: 4 percent of the outstanding stock.
  - Characteristics of outstanding bonds without CACs:
    - About 44 percent of outstanding bonds without any CACs are below investment-grade.
    - Outstanding bonds without CACs would not mature until 2096.
    - 75 percent of outstanding bonds without CACs are under New York law.
  - Pari passu clause: about 50 percent of outstanding stock includes the ICMA’s strengthened pari passu clause.
- Data sources and estimation notes:
  - Share of outstanding stock calculated based on outstanding amount in US dollars as of end-March 2019 (data used in IMF’s Fourth Progress Report on Inclusion of Enhanced Contractual Provisions in International Sovereign Bond Contracts (2019)).
  - Data sources mentioned: Bloomberg, Dealogic, Perfect Information database, various countries’ authorities; excludes GDP warrants and China’s domestic issuances under Hong Kong SAR governing law.
  - Figure and chart references in source: Figure 1 “Foreign Law-Governed Sovereign Bonds: Overview (as of end-March 2020)”; Figure 2 “Total International Portfolio Investment Assets (in trillions of US dollars; as of June 2019)”.
    - Figure 2 listed values (as presented): 12.5, 4.8, 4.5, 3.6, 3.5, 3.4, 3.0, 2.4, 2.1, 1.8 (units: trillions of US dollars; as of June 2019).

### IV. Empirical approach and principal findings (summary statements from text)
- Data use: primarily emerging-market economies’ bond yield spreads (secondary-market) over May 1996–March 2020; some primary-market yield observations used to assess issuance cost.
- Core empirical findings (as stated):
  - Regular CACs: associated with lower borrowing costs most of the time in secondary-market yields.
  - Enhanced CACs (ICMA model, post-August 2014): associated with lower spreads since introduction.
  - Market interpretation: investors consider benefits of orderly and efficient restructuring rather than attributing increased moral hazard to CACs or enhanced CACs.
- Contextual inference: results qualify earlier IMF Progress Reports’ conclusion that enhanced CACs have no observable pricing effect by showing a post-August 2014 association with lower spreads in secondary-market data.

### V. Observations on crisis vs non-crisis periods (literature and empirical motivations)
- Prior literature findings summarized:
  - Some studies find CAC effects differ by credit rating (mid-rated issuers may benefit most).
  - During distress, evidence is mixed: in some cases CACs associated with higher yields in near-default situations (example: Venezuela 2010–16).
- Paper’s stated objective: examine yield developments at times of debt distress which are key to restructurings and exchanges, using secondary-market pricing to capture behavior at particular distress points.

### VI. Organization of paper and annex references
- Structure as described:
  - Section II: brief literature review.
  - Section III: overview of evolution of CACs and current status.
  - Sections IV and V: empirical analysis findings for different historical periods (including pre-crisis, Global Financial Crisis, European Sovereign Debt Crisis, and ICMA enhanced CACs regressions, as indicated by table listings).
  - Section VI: conclusions and insights on interpreting findings.
- Annexes listed in source:
  - Annex I: Description of Variables.
  - Annex II: Alternative Regression Results to Table 7.

*Source: wpiea2020162-print-pdf - Annex I : Description of Variables (excerpt provided).*

### Conclusion of EFSM/EFSF program

### Conclusion of EFSM/EFSF program

### Data and sample
- Sample: 1,025 outstanding foreign law-governed sovereign bonds with secondary-market yields available at end of March 2020.
- Time coverage: monthly series from May 1996 to March 2020 (287 time-points).
- Countries: bonds from 116 countries for all regions (advanced economies and emerging markets); emerging markets broadly defined, including frontier markets and low-income countries.
- Bond selection rules:
  - Omitted bonds with remaining maturity of less than one year.
  - Excluded spreads over 1,000.
  - Included mostly conventional-type bonds (fixed rate, bullet payment, or simple coupon); excluded complex coupon structures, convertibles, or variable rates.
  - Treated central bank bonds issued to finance the sovereign balance sheet as equivalent to government bonds issued by the ministry of finance.
  - Excluded state-owned enterprise bonds and government guaranteed bonds.
  - Included sukuk issued in international markets using their regular daily yields.
- CAC coverage in sample: 4 percent bonds with no CACs, 45 percent with regular CACs, and 51 percent with enhanced CACs.
- Primary pricing sources: BVAL and CBBT; fallback to Bloomberg generic BGN or others.
- Data sources: Bloomberg, IMF World Economic Outlook (WEO) database, Perfect Information database, Dealogic.
- Legal indexing: detailed indexing of legal clauses (various forms of CACs, enhanced CACs, pari passu and strengthened pari passu clauses) developed for this study.

### Methodology
- Dependent variable: sovereign bond yield spread (basis points) over relevant currency benchmarks (10-year U.S. government treasury, German bund, Japanese government bond yields).
- Mid-yields to maturity: simple average of daily series, then converted to spreads versus corresponding benchmarks.
- Regression framework: panel regression model with bond-specific time-invariant effects.
  - Specification: Yi,t = α + β1 CAC + β2 CAC_ENHANCED + β3 Xi,t + β4 θi + εi,t
  - CAC variables: three mutually exclusive binary dummies — no CACs (1 or 0), regular CACs (1 or 0), enhanced CACs (1 or 0).
  - Interpretation example: a regular CAC coefficient of -40 implies the yield of bonds with CACs is 40 basis points lower than bonds without CACs.
- Independent variables (Xi,t) include:
  - Bond-specific: coupon, tranche volume in billions of US dollars, tenor, governing law (binary), SEC registration, stock exchange, issuing currency, origin maturity.
  - Time-variant macro/market: inflation (annual percent changes), sovereign five-year CDSs, changes in numeric credit ratings, remaining maturities (years), bid-ask spreads, VIX, exchange rates, debt-to-GDP ratios, EMBI spreads.
- Estimation details:
  - Monthly series used to reduce daily-yield noise.
  - Models run for whole sample and for subperiods; omitted variables when encountering near singular matrices.
  - Credit rating coding: numeric conversion where “lower the grade, the larger the numbers” (example: AAA coded 1, C coded 24).

### Key empirical findings on CACs
- Overall effect:
  - The presence of CACs is associated with lower secondary-market sovereign bond yield spreads.
  - Relationship is more pronounced for noninvestment-grade issuers.
- Regular CACs:
  - Most of the time, regular CACs are negatively associated with bond yield spreads, except during the European sovereign debt crisis period.
  - For noninvestment-grade issues, the negative association with spreads is more consistent.
  - For investment-grade issues, regular CACs have a positive coefficient (associated with higher borrowing costs) across the considered times.
- Enhanced CACs (ICMA introduction in August 2014):
  - Bonds that included enhanced CACs exhibit negative coefficients with spreads.
  - Presence of enhanced CACs is associated with lower secondary-market yield spreads and these results are consistently statistically significant.
  - Findings suggest investors do not price bonds adversely based solely on inclusion of enhanced CACs; results likely to persist after current market distress.

### Determinants of secondary-market yield spreads (selected results)
- Credit ratings (time-variant):
  - Strong explanatory variable. More creditworthy bonds (numeric values 1–10) associated with lower yield spreads; less creditworthy bonds (numeric values over 10) associated with higher yield spreads.
- Remaining maturities:
  - Positively associated with spreads; more months to maturity imply higher yield spreads.
- Bond-specific characteristics:
  - Coupon and tranche volumes: positively associated with yield spreads; tranche volumes important for yields.
  - Issue currency: issuing in euros or US dollars associated with lower yield spreads compared to Japanese yen, British pound, Scandinavian currencies, Chinese renminbi.
  - Governing law: not consistently significant over full sample; English governing law associated with higher yield spreads only in the latest periods (after 2014) and statistically significant then.
- Liquidity (bid-ask spreads) and SEC registration:
  - Bid-ask spreads positively associated with yield spreads (i.e., greater liquidity — lower bid-ask spread — associated with lower yield spreads).
  - SEC registration shows a statistically strong negative relationship with yield spreads since 2010 (more liquid/transparent bonds associated with lower spreads).
- Macroeconomic and financial market variables:
  - Consumer price index annual growth rate: positively associated with yield spreads.
  - CDS spread and market VIX: positively associated with yield spreads.
  - Commodity prices: not statistically significant, mixed signs.
  - Monetary union membership: associated with higher secondary-market bond yield spreads during the European sovereign debt crisis.
- General conclusion: macroeconomic fundamentals and market variables play a major role in explaining government bond yield differentials.

### Novelty and scope of study
- Novel dataset and indexing of legal clauses allow systematic analysis of effects of inclusion of CACs and enhanced CACs on secondary-market yields.
- Comprehensive coverage: 1,025 bonds, 287 monthly observations, 116 countries, advanced and emerging markets, multiple issuing currencies, governing laws, market liquidity measures, credit ratings, and macroeconomic indicators — providing broader scope than prior studies focused on specific markets, periods, or governing laws.

*Source: Conclusion of EFSM/EFSF program (excerpt from the provided PDF content).*

### 1. Before New York law-governed sovereign bonds included CACs: May 1996–

### 1. Before New York law-governed sovereign bonds included CACs: May 1996–

### Periodization and data treatment
- Study periods:
  - Before New York law-governed sovereign bonds included CACs (May 1996–January 2003)
  - Pre-global financial crisis (February 2003–May 2007)
  - Global financial crisis (June 2007–December 2009)
  - European sovereign debt crisis (January 2010–August 2014)
  - After ICMA introduced enhanced CACs (September 2014–March 2020)
- Venezuela pricing data: excluded (Bloomberg pricing data for 14 Venezuela bonds) for the period September 2014–March 2020 because of highly volatile secondary-market trading activity that distorted results.
- Methodological note: For the earliest period (pre-February 2003) limited data induced a near singular matrix problem, requiring reduction in number of variables for model fit.

### Empirical findings — secondary-market bond yield spreads (by period)
- Before New York law-governed sovereign bonds included CACs (pre-February 2003)
  - CACs are estimated to have contributed to lowering yield spreads by 1.3 percent (statistically significant) for investment grade issuers.
  - Effect of CACs is insignificant for noninvestment-grade issuers.
  - (Results summarized in Table 3 of the source.)

- Pre-global financial crisis (February 2003–May 2007)
  - After New York law-governed sovereign bonds prevalently included CACs, spreads of bonds with CACs are lower than those without CACs in the entire sample.
  - Evidence that CACs lead to a decrease in secondary-market bond yield spreads and borrowing costs for both noninvestment-grade and investment-grade issuers (all results statistically significant).
  - Selected regression indicators (Table 4): R-squared values reported as 0.42 (Total), 0.34 (Non-investment grade), 0.66 (Investment grade).
  - Notes: **99% significant, * 95% significant.

- Global financial crisis (June 2007–December 2009)
  - Inclusion of CACs appears to decrease bond yield spreads for noninvestment-grade countries by 402 basis points (statistically significant).
  - For investment-grade issuers, CACs are positively associated with spreads (statistically significant).
  - R-squared values reported as 0.97 (Total), 0.89 (Non-investment grade), 0.90 (Investment grade).
  - Notes: **99% significant, * 95% significant.

- European sovereign debt crisis (January 2010–August 2014)
  - For the overall sample inclusion of CACs seems associated with slightly higher borrowing costs, with a statistically significant impact.
  - For both investment-grade and noninvestment-grade issuers, inclusion of CACs is associated with higher bond yield spreads (effects statistically significant). Results are affected by Greece’s inclusion (noted S&P rating movements between 2010 and 2014).
  - Reported R-squared: 0.87 (Total), 0.91 (Non-investment grade), 0.76 (Investment grade).
  - Notes: **99% significant, * 95% significant.

- After ICMA introduced enhanced CACs (September 2014–March 2020)
  - Presence of enhanced (single-limb) CACs is negatively associated (statistically significant) with bond yield spreads for both noninvestment-grade and investment-grade issuers (Table 7).
  - Regular CACs also show a negative and statistically significant association with bond spreads for the whole sample.
  - Empirical implication: inclusion of enhanced CACs and regular CACs is associated with lower borrowing costs for the sovereign.
  - Reported regression highlights (Table 7):
    - Enhanced CACs coefficient: -17.9 (Total) **, -129.8 (Non-investment grade) **, -16.6 * (Investment grade).
    - Regular CACs coefficient: -31.9 ** (Total), -85.9 ** (Non-investment grade), -34.5 ** (Investment grade).
    - R-squared values: 0.73 (Total), 0.45 (Non-investment grade), 0.64 (Investment grade).
  - Notes: **99% significant, * 95% significant.

### Primary-market observations for selected countries
- Case selection: Mexico (New York law, investment-grade), Romania (UK law, investment-grade), Indonesia (New York law, noninvestment-grade) chosen for similar bond characteristics (coupon, remaining maturity, tranche volume, currency, ratings).
- Finding: Yields of bonds with enhanced CACs at issuance generally align with the sovereign yield curve of the issuer; no obvious pricing premium for enhanced CACs at issuance was observed.
  - Mexico’s bonds with enhanced CACs priced consistently with Mexico’s sovereign USD yield curve.
  - Romania’s new EUR bonds with enhanced CACs priced aligned with the EUR sovereign yield curve during normal and distressed times.
  - No material pricing difference for enhanced-CAC bonds observed in secondary market during market lows and highs for these issuers.

### Conclusion and interpretation
- Overall result: Using an extensive secondary-market bond spread dataset, the study finds that:
  - Regular CACs lowered the cost of borrowing for most considered historical periods, with effects more pronounced for noninvestment-grade issuers.
  - The bond-yield-spread/borrowing-cost-reducing effects are strongest in the first two periods (pre-2003 and 2003–2007) and lose statistical significance starting with the global financial crisis.
  - Enhanced (single-limb) CACs are associated with lower bond yield spreads.
- Interpretation: Findings consistent with secondary-market investors’ implied acknowledgement of CACs’ potential benefits to an efficient and orderly restructuring process, helping especially the pricing of noninvestment-grade issues.
- Caveats and limitations noted for further investigation:
  - Composition of the investor base (can change quickly and affect pricing; institutional investors may be replaced by hedge funds or official creditors during distress).
  - Variations in CAC formulations (minimum voting requirements of 66.6 percent, 75 percent, 85 percent, and 100 percent; mandatory meeting requirements; aggregation; acceleration; reverse acceleration; collective representation) not distinguished in this study.
  - Other contractual provisions (pari passu/modified pari passu, cross-default clauses, engagement clauses, disenfranchised clauses, information covenant clauses) may influence pricing and were not separately analyzed.
  - Variations within rating groups (e.g., within investment-grade or within BBB+/BBB/BBB-) and issuer structures (trustee vs fiscal agency) were not differentiated.
  - Treatment of restructured/exchanged bonds (different payment schedules, amortization) may affect pricing and warrants further study.

*Source: IMF working paper chapter (regression tables and text excerpts as provided in the supplied content).*

### 237. Washington: International Monetary Fund.

### wpiea2020162-print-pdf - 237. Washington: International Monetary Fund.

### References and literature cited
- Committee on International Economic Policy and Reform (CIEPR), 2013. “Revisiting Sovereign Bankruptcy,” October. Washington: Brookings Institution.
- Cruces, Juan J., and Christoph Trebesch, 2013. “Sovereign Defaults: The Price of Haircuts,” American Economic Journal: Macroeconomics, 5(3), 85–117.
- Das, Udaibir S., Michael G. Papaioannou, and Christoph Trebesch, 2013. “Sovereign Debt Restructurings 1950–2010: Literature Survey, Data, and Stylized Facts,” IMF Working Paper 12/203. Washington: International Monetary Fund.
- Das, Udaibir S., Michael G. Papaioannou, and Christoph Trebesch, 2010. “Sovereign Default Risk and Private Sector Access to Capital in Emerging Markets,” IMF Working Paper 10/10. Washington: International Monetary Fund.
- De Grauwe, Paul, 2011. “The Governance of a Fragile Eurozone,” Centre for European Policy Studies.
- Döemeland, Döerte (ed), 2009. “Debt Relief and Beyond: Lessons Learned and Challenges Ahead,” Chapter 7, 141–79. Washington: World Bank.
- Duggar, Elena, 2013. “Investor Losses in Modern-Era Sovereign Bond Restructurings,” Moody’s Sovereign Defaults Series, October 7.
- Duggar, Elena, 2013. “The Role of Holdout Creditors and CACs in Sovereign Debt Restructurings,” Moody’s Sovereign Defaults Series, October 7.
- EFC Sub-Committee on EU Sovereign Debt Markets, 2012. “Model Collective Action Clause,” Supplemental Explanatory Note, March 26. Brussels: EU Economic and Financial Committee (EFC).
- Eichengreen, Barry, and Ashoka Mody, 2004. “Do Collective Action Clauses Raise Borrowing Costs?” Economic Journal, 114(495): 247–64.
- Eichengreen, Barry, and Ashoka Mody, 2003. “Is Aggregation a Problem for Sovereign Debt Restructuring?” American Economic Review, 93(2), 80–4.
- Eichengreen, Barry, and Ashoka Mody, 2000. “Would Collective Action Clauses Raise Borrowing Costs?” NBER Working Paper 7458.
- Enderlein, Henrik, Christoph Trebesch, and Laura von Daniels, 2012. “Sovereign Debt Disputes: A Database on Government Coerciveness during Debt Crises,” Journal of International Money and Finance, March 31(2), 250–66.
- Eurogroup, 2010. Statement by the Eurogroup, November 28, 2010.
- Eurozone, 2012. Treaty Establishing the European Stability Mechanism.
- Fang, Chuck, Julian Schumacher, and Christoph Trebesch, 2019. “Restructuring Sovereign Bonds: Holdouts, Haircuts and Effectiveness of CACs,” paper presented at the 20th Jacque Pollak Annual Research Conference, Washington: International Monetary Fund.
- Fang, Chuck, 2015. “Cost and efficacy of collective action clauses,” Institute of International Studies, University of California, Berkeley.
- Fernandez-Ansola, Juan Jose, and Thomas Laursen, 1995. “Historical Experience with Bond Financing,” IMF Working Paper 95/27. Washington: International Monetary Fund.
- Gadanecz, Blaise, 2004. “The Syndicated Loan Market: Structure, Development and Implications.” BIS Quarterly Review, December. Basel: Bank of International Settlements.
- Gelpern, Anna, 2013. “A Skeptic’s Case for Sovereign Bankruptcy,” Houston Law Review, 2013, Vol. 50, 1095–1127.
- Gelpern, Anna, 2003. “How collective action is changing sovereign debt,” International Financial Law Review 22, 19–22.
- Gelpern, Anna, and Mitu Gulati, 2013. “The Wonder-Clause,” Journal of Comparative Economics, 41(2), 311–652.
- Gelpern, Anna, and Mitu Gulati, 2009. “Innovation after the revolution: foreign sovereign bond contracts after 2003,” Capital Markets Law Journal 4, 85–103.
- Gelpern, Anna, and Mitu Gulati, 2006. “Public symbol in private contract: a case study,” Washington University Law Quarterly 84, 1627–1715.
- Government of Belize, 2007. Belize 2015 Notes Amended, Press Release, February 5, 2007.
- Group of Ten, 1996. “Resolving Sovereign Liquidity Crises.” Washington: Group of Ten.
- The Guardian, 2011. “Vulture Funds: The Key Players,” November 15, 2011.
- Häseler, Sönke, 2012. “Trustees versus Fiscal Agents and Default Risk in International Sovereign Bonds,” European Journal of Law and Economics.
- Häseler, Sönke, 2009. “Collective Action Clauses in International Sovereign Bond Contracts—Whence the Opposition?” Journal of Economic Surveys (2009), Vol. 23, No. 5, 882–923.
- International Monetary Fund, 2019. “Fourth Progress Report on Inclusion of Enhanced Contractual Provisions in International Sovereign Bond Contracts,” Policy Paper. March. Washington: International Monetary Fund.
- International Monetary Fund, 2017. “Third Progress Report on Inclusion of Enhanced Contractual Provisions in International Sovereign Bond Contracts,” Policy Paper, December. Washington: International Monetary Fund.
- International Monetary Fund, 2016. “Second Progress Report on Inclusion of Enhanced Contractual Provisions in International Sovereign Bond Contracts,” Policy Paper, December. Washington: International Monetary Fund.
- International Monetary Fund, 2014. “Strengthening the Contractual Framework to Address the Collective Action Problems in Sovereign Debt Restructuring,” Policy Paper, October. Washington: International Monetary Fund.
- International Monetary Fund, 2012. St. Kitts and Nevis, IMF Country Report 12/284. Washington: International Monetary Fund.
- International Monetary Fund, 2003. “Proposals for a Sovereign Debt Restructuring Mechanism (SDRM). Washington: International Monetary Fund.
- Muse-Fisher, John, 2014. “Starving the Vultures: NML Capital v. Republic of Argentina and Solutions to the Problem of Distressed-Debt Funds,” California Law Review, 2014, 120(6), Article 8, 1671–1726.
- Olivares-Caminal, Rodrigo, 2013. “The Pari passu Clause in Sovereign Debt Instruments: Developments in Recent Litigation,” BIS Papers No. 72, 121–128.
- Panizza, Ugo, Federico Sturzenegger, and Jeromin Zettelmeyer, 2009. “The Economics and Law of Sovereign Debt and Default,” Journal of Economic Literature, 47(3), 651–98.
- Paulus, Christoph, 2010. “A Standing Arbitral Tribunal as a Procedural Solution for Sovereign Debt Restructurings,” in Sovereign Debt and the Financial Crisis: Will This Time Be Different, Carlos Braga and Gall Vincelette, eds, 317-29. Washington: World Bank.
- Picarelli, Mattia Osvaldo, Aitor Erce, and Xu Jiang, 2019. “The Benefits of Reducing Holdout Risk: Evidence from the Euro CAC Experiment, 2013–2018,” Capital Markets Law Journal, 14(2), April, 155–177, https://doi.org/10.1093/cmlj/kmz005.
- Ratha, Dilip, Supriyo De, and Sergio Kurlat, 2016. “Does Governing Law Affect Bond Spreads?” Policy Research Working Paper, No. 7863. Washington: World Bank.
- Reinhart, Carmen M., and Christoph Trebesch, 2014. “A Distant Mirror of Debt, Default, and Relief,” NBER Working Paper 20577.
- Remolona, Eli M, Michela Scatigna, and Eliza Wu, 2007. “Interpreting sovereign spreads”, BIS Quarterly Review, March 2007.
- Richards, Anthony J., and Gugiatti, Mark, 2003. “Do Collective Action Clauses Influence Bond Yields? New Evidence from Emerging Markets,” International Finance, 6(3), 415-47.
- Rogoff, Kenneth, and Jeromin Zettelmeyer, 2002. “Bankruptcy Procedures for Sovereigns: A History of Ideas, 1976-2001,” IMF Staff Paper, 49(3), 470–507. Washington: International Monetary Fund.
- Schumacher, Julian, Christoph Trebesch, and Henrik Enderlein, 2018. “Sovereign Defaults in Court,” ECB Working Paper Series No. 2135, February. Frankfurt am Main: European Central Bank.
- Steffen, Christoph Grosse, Sebastian Grund, and Julian Schumacher, 2019. “Collective Action Clauses in the Euro Area: A Law and Economic Analysis of the First Five Years,” Capital Markets Law Journal, 14(2), April, 134–154, https://doi.org/10.1093/cmlj/kmz001.
- Stolper, Antonia E., and Sean Dougherty, 2017. “Collective Action Clauses: How the Argentina Litigation Changed the Sovereign Debt Markets,” Capital Markets Law Journal, 12(2), April, 239–252, https://doi.org/10.1093/cmlj/kmx022.
- Sturzenegger, Federico, and Jeromin Zettelmeyer, 2005. “Haircuts: Estimating Investor Losses in Sovereign Debt Restructurings, 1998–2005,” IMF Working Paper 05/137. Washington: International Monetary Fund.
- Trebesch, Christoph, 2008. “Delays in Sovereign Debt Restructurings: Should We Really Blame the Creditors?” Proceedings of the German Development Economics Conference, Zurich 2008. 44 Vernein Fur socialpolitik, Research Committee Development Economics.
- Weidemaier, Mark C., and Mitu Gulati, 2012. “A People’s History of Collective Action Clauses,” Virginia Journal of International Law, 2013, 54(1), 52–95.
- Zettelmeyer, Jeromin, Christoph Trebesch, and Mitu Gulati, 2013. “The Greek Debt Exchange: An Autopsy,” Economic Policy, 2013, 28(75), 513–56.

### Annex I: Description of variables (as used in primary- and secondary-market analyses)
- Mid-yield to maturity: Monthly data are based on the simple average of daily mid-yield to maturity from May 1996 to March 2020 (time-variant).
- Sovereign bond yield spread over relevant benchmark: Calculated based on the issue currency: US dollar → US treasury 10-year generic yields; euro → German 10-year bond generic yield; Yen → Japanese 10-year bond generic yields.
- Credit rating: S&P foreign-currency long-term sovereign credit ratings are converted to numeric values, as outlined in Annex II Table 2. Complemented by Moody’s and Fitch when S&P rating is not available. Converted to a numeric value over the months (time-variant).
- Credit rating at issuance: S&P foreign-currency long-term sovereign credit ratings at issuance are converted to numeric value, as outlined in Annex II, Table 2.
- Maturity: year of maturity.
- Remaining maturities: Years to maturity is calculated based on the last day of the month (time-variant).
- Bid-ask spread: Calculated based on bid and ask prices for each month. Monthly series is based on the average of daily series whenever there was a meaningful bid price and ask price (time-variant).
- Coupon: Each bond’s coupon in percent (time-invariant).
- Outstanding $ amount: Each bond’s outstanding tranche volume, not the total deal volume. Each tranche volume is converted to billions of US dollars based on the exchange rate of the date of issuance (time-invariant).
- EUR currency: If a bond is issued in euros, the value is 1, otherwise 0 (binary value).
- USD currency: If a bond is issued in US dollars, the value is 1, otherwise 0.
- English law: If the governing law is English law, the value is 1, otherwise 0, with the significant majority of other cases being New York law. Also, another governing-law category is employed; English law, New York law, and other governing law are mutually exclusive and their value adds up to 1 for each bond. Other governing law comprises less than 2 percent of total cases (time-invariant).
- CDS spread: Monthly series of each sovereign issuer’s CDS spread is used for each specific bond. Monthly series are simple averages of daily series (time-invariant).
- VIX: CBOE Volatility Index at the time of the issue date, measure of implied volatility of S&P 500 index options, calculated and published by the CBOE (time-variant).
- SEC: If a bond is registered in SEC at the time of issuance, the value is 1, otherwise 0.
- Consumer Price Index (CPI): Annualized consumer inflation growth rate in percent, monthly.
- Exchange rate (EXR): National currency to US dollar, monthly average of daily exchange rates.
- Debt-to-GDP ratio of previous year: Debt to GDP in percent in the year t-1.
- Commodity price index.
- Monetary union: If the issuer is a member of a monetary union at each data point, the value is 1, otherwise 0.
- No CACs: Based on sales documents and prospectuses from Perfect Information, Dealogic, and Bloomberg, if a bond does not include collective action clauses, the value is 1, otherwise 0. Up to August 2014, no CACs and CACs are mutually exclusive and add up to 1 for each bond. For September 2014 to November 2016, no CACs, CACs, and enhanced CACs are mutually exclusive and add up to 1 for each bond.
- Regular CACs: Based on sales documents and prospectuses from Perfect Information, Dealogic, and Bloomberg, if a bond includes collective action clauses, the value is 1, otherwise 0 (time-invariant).
- Enhanced CACs: Based on sales documents and prospectuses from Perfect Information, Dealogic, and Bloomberg, if a bond includes ICMA’s enhanced version of collective action clauses, the value is 1, otherwise 0. IMF Legal Department staff verify the correct indexing of this information (time-invariant).
- Emerging markets (EM): If the issuing country is an Emerging Market and not an advanced economy, according to the IMF WEO definition, the value is 1, otherwise 0 (binary value).

### Annex II: Alternative regression results (summary and key statistics)
- Purpose: Limit sample to U.S. dollar-denominated sovereign bond issuances; eliminate potentially endogenous independent variables (time-variant sovereign credit ratings, sovereign five-year credit default swap, bid-ask spread); exclude spreads over 1,000 basis points. Period: September 2014–March 2020.
- Main inference from the revised model:
  - Inclusion of single-limb CACs (ICMA enhanced CACs) is associated with lower borrowing costs for both noninvestment- and investment-grade issuers.
  - Inclusion of regular CACs is associated with even lower borrowing costs.
  - Noninvestment-grade issuers benefit more from including single-limb CACs than investment-grade issuers.
  - Including bonds with spreads over 1,000 basis points yields similar coefficient results but R-squared falls drastically to 22 percent.
  - Elimination of potentially endogenous variables does not necessarily lead to but supports the causal effects between dependent and independent variables.
  - Results indicate market participants do not associate inclusion of enhanced CACs with potential future restructurings.

- Secondary markets: Panel regression results
  - Dependent variable: Spread over benchmark in bps
  - Coefficient estimates (Total | Non-investment grade | Investment grade):
    - Coupon: 27.7 ** | -5.9 | 11.8 **
    - Outstanding $: -52.4 | -111.2 | 19.9 **
    - Maturity: 3.4 ** | 4.5 ** | 2.6 **
    - NY law: -41.4 ** | -160.0 ** | -18.7 **
    - SEC: -15.4 ** | 42.7 ** | -42.0 **
    - EMBI: 0.5 ** | 0.4 ** | 0.3 **
    - VIX: 1.1 ** | 2.3 ** | 1.4 **
    - Exchange rate: 0.0 ** | 0.0 | 0.0 **
    - CPI(-6): 0.1 | -1.5 | -3.3
    - DEBT(-6): 0.3 * | 1.6 * | -1.6 *
    - Enhanced CACs: -29.6 * | -70.5 * | -35.4 **
    - Regular CACs: -59.9 ** | -113.7 ** | -28.9 **
    - C (constant): -676.0 | -869.2 | -515.9
    - R-squared: 0.66 | 0.48 | 0.69
  - Notes: **99% significant, * 95% significant.
  - Context: ICMA introduced enhanced CACs; sample period 2014M9-2020M3.

### Annex II: Conversion of credit ratings to numeric values (Annex II. Table 2)
- Composite credit ratings → Numeric value:
  - AAA → 1
  - AA+ → 2
  - AA → 3
  - AA- → 4
  - A+ → 5
  - A → 6
  - A- → 7
  - BBB+ → 8
  - BBB → 9
  - BBB- → 10
  - BB+ → 11
  - BB → 12
  - BB- → 13
  - B+ → 14
  - B → 15
  - B- → 16
  - CCC+ → 17
  - CCC → 18
  - CCC- → 19
  - CC+ → 20
  - CC → 21
  - CC- → 22
  - C+ → 23
  - C → 24
  - C- → 25
  - D → 26
  - Not rated → 27

*Source: wpiea2020162-print-pdf - 237. Washington: International Monetary Fund.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020162-print-pdf.pdf_