## 2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts ## Source details **Canonical URL:** [2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts](https://www.imf.org/-/media/files/publications/wp/2018/wp18263.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2018/wp18263.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2018/wp18263.pdf.json) --- ### Data and sample - Sample period: 1994-2014. - Total syndicated loans after cleaning: 16,847 syndicated loans. - Borrowers in sample: 7,589 borrowers headquartered in 107 emerging and developing countries. - Pricing subsample: at most 7,571 deals (and 3,703 borrowers). - Sample dominated by large emerging markets (China, Brazil, India, Mexico, Indonesia, and Turkey); borrowers from low-income and lower middle-income countries represent more than 30 percent of the sample. - Loans are cross-border syndicated loans (borrower and lenders from different countries); sovereign loans excluded. - Loan size measured in 2011 constant USD and excludes domestic-bank-financed portion. ### Loan pricing, maturity, and size — key statistics - Baseline pricing measure: all-in interest rate spread (contract spread over LIBOR plus any annual fee and any upfront fee). - Average all-in interest rate spread: 351 bps. - Interquantile range of all-in spread: 180 to 475 bps. - Median loan maturity: 3-year maturity. - Share of loans with maturity of one year or shorter: 27 percent. - Share of loans longer than 10 years: 10 percent. - Median loan size: USD 65 million. - Lower quartile loan size: USD 21 million. - Upper quartile loan size: USD 170 million. ### Forms of MDB participation and prevalence - MDB participation takes two forms: A/B loans and parallel loans. - A/B loans: MDB is lender of record (A Loan) and invites external participants for the B Loan. - Parallel loans: MDB and external sources each conclude separate loan agreements with borrower on MDB-administered project. - Share of deals with at least one MDB in the syndicate: about 10 percent (1,694 deals). - Share of MDB-involved deals that are A/B loans: 63 percent. - Main MDBs in sample: European Bank for Reconstruction and Development and International Finance Corporation together account for 56 percent of MDB-involved deals; other key players include the European Investment Bank, the International Bank for Reconstruction and Development, the African Development Bank, the Inter-American Development Bank, and the Asian Development Bank. ### Stylized comparisons: loans with vs. without MDB participation - Average all-in spread premium for loans with MDB participation vs. private-only syndicates: 96 bps more. - Premium reflects an almost equal difference in interest rate spread and in fees. - Average maturity difference: loans with MDB participation have 32 months longer maturity. - Average size difference: deals with MDB participation are smaller by about USD 28 million. - Industry concentration: MDB participation is widespread across industries with a concentration in agriculture and lower presence in manufacturing and natural resources. - Geographic concentration: MDB participation is more common in lending to low- and lower middle-income countries than to emerging markets. - Interpretation: higher spreads and longer maturities are consistent with MDBs’ capacity to lend at longer tenure and with a higher propensity to finance riskier projects (especially infrastructure) that private sector may avoid. ### De-risking evidence and interpretation - Presence of an MDB in a syndicate is associated with a 37 percent reduction in the premium paid by risky borrowers, suggesting a de-risking role (lowering borrowing costs for risky firms). - Possible channels for de-risking: - Informational advantages and strong monitoring capacity of MDBs. - Extension of MDBs’ de facto preferred creditor status to syndicate participants (noted especially in A/B arrangements). - Results control for deal characteristics and absorb time-varying unobserved heterogeneity at industry and country level, including country×industry fixed effects; results robust to sub-samples and matching techniques. ### Macro trends in cross-border syndicated lending and MDB role - Cross-border syndicated loan flows show a cyclical trend: increasing in early 1990s, decline in early 2000s, rapid surge until the global financial crisis, then slight decline. - MDB participation follows a similar pattern and becomes relatively more important during downward phases of the cycle (early 2000s and post-global financial crisis), consistent with a counter-cyclical role. - During downward phases, loans with MDB participation amounted to up to 15 percent of all cross-border lending. - In more recent years (within sample), this share declined to below 10 percent. - Regional shift since 2007: increasing importance of cross-border syndicated lending to low-income countries, especially in South Asia and Sub-Saharan Africa; lending to low-income countries has a strong infrastructure financing component but remains concentrated in a few recipients. --- ### 3.2 Loan Pricing — empirical specification and explanatory power - Table 3 estimates equation 1 with the all-in spread as dependent variable. Column 1 to 5 incrementally add fixed effects up to the preferred specification with country×year and industry×year fixed effects; column 6 includes country×industry fixed effects. - The comparison of the R2 across specifications: - R2 increases from 0.38 to 0.51 between column 1 and 2, indicating global shocks (year fixed effects) play a key role. - Time-varying country and industry fixed effects raise the R2 to 0.66 (column 5), suggesting the model captures two third of the observed variation in loan prices across borrowers. ### MDB participation and pricing (regression evidence) - The coefficient on the MDBs’ participation dummy is always positive and statistically significant across specifications, ranging from 82 (column 2, with country and year fixed effects) to 45 (column 5, with country×year and industry×year fixed effects). - Preferred/conservative specification (column 5) implies: - Loans with MDBs’ participation are more expensive by 45 bps or 13 percent (relative to the average all-in spread of 351 bps). - Interpretation: MDBs may self-select into loans with higher risk and higher spreads that private sector would not finance. ### Deal characteristics and pricing - Maturity: an additional year of maturity is associated with a 9 bps increase in the all-in spread (column 5). - Loan size and complexity: - Smaller loans and those with longer maturity are associated with higher prices. - A higher number of tranches is associated with higher prices. - Borrower credit risk: - Highly leveraged borrowers pay on average 365 bps more than investment grade borrowers (column 5). - Leveraged borrowers pay a premium of 115 bps (column 5). - Currency: Loans in Euro and in other currencies have a discount compared with loans in USD. - Loan type: Term loans cost 37 bps more than a credit facility (column 5). - Borrower sector: Public sector companies and government pay 55 bps less than private sector ones. - Guarantor: Deals with a guarantor do not show statistical difference in price from loans without guarantor. - Syndicate concentration (HHI): One standard deviation in the HHI is associated with a discount of 26 bps (column 5). ### Spread versus fees - Results hold when using spread and fees separately (Table 4). - Premium due to MDBs’ presence is almost equally split: - Higher fees: 19 bps. - Higher loan spreads: 25 bps. - Maturity, term loans, and syndicate concentration have larger effects on fees than on spreads. --- ### 3.3 De-Risking — hypothesis, identification, and evidence - Hypothesis: MDBs may participate when private market cannot provide funding due to high borrower risk; they also may reduce borrowing costs through de-risking measures (informational advantages, better monitoring, de facto senior creditor status). - Two proxies for borrower creditworthiness used in interactions with the MDB participation dummy: - Risky: dummy = 1 for leveraged or highly leveraged deals. - High country risk: dummy = 1 for borrowers in countries in the bottom half of the Institutional Investor country credit rating distribution. ### Descriptive and regression evidence - Descriptive: For deals without MDBs, strong association between borrower riskiness and all-in spread; with MDBs present, risky borrowers obtain loans priced similar to less risky borrowers. - Table 5 regression results (country×year and industry×year fixed effects): - MDBs’ participation reduces the premium associated with leveraged and highly leveraged borrowers by about a third (column 1). - De-risking effect remains when considering spread and fees separately; effect larger and more precisely estimated for spreads (columns 3 and 5). - Presence of an MDB is associated with significantly lower borrowing costs (about 41 bps) for companies headquartered in riskier countries, controlling for deal characteristics including borrower creditworthiness. --- ### 3.4 Infrastructure and Public Sector Lending - Interaction of MDB participation with infrastructure project indicator yields a positive and significant coefficient: - Infrastructure loans with MDBs’ participation are about 66 bps more expensive than similar loans financed entirely by commercial banks (Table 6, column 1). - This premium is mostly driven by changes in spread rather than fees (columns 3 and 5). - Interpretation: MDBs tend to finance infrastructure projects with higher risks compared to similar projects financed by commercial banks alone. - Public versus private sector borrowers: - MDB participation is associated with significantly lower borrowing costs for public sector firms. - Economically sizable effect: presence of an MDB in the syndicate almost doubles the reduction in the all-in spread for public sector borrowers compared to private sector ones (Table 6, column 2). - Effect is almost equally large across spread and fees, though the fee estimate is not statistically significant (column 6). --- ### 3.5 Other Loan Terms — size and maturity - Loan size: - MDB participation is associated with lower loan size. - Column 2 estimate: deals with MDBs are on average USD 70 million smaller than loans granted only by commercial banks—almost 40 percent smaller than the average loan of about USD 180 million. - Other findings: - Highly leveraged borrowers obtain smaller loans. - Longer and more complex loans have larger loan size. - Term loans are generally smaller than credit facilities by around USD 35 million. - Higher lender concentration is associated with smaller deals. - Maturity: - MDB participation is associated with longer maturities: - Difference ranges from about 25 months to 27 months across specifications (columns 4-6). - Baseline (column 5): loans with an MDB are on average 27 months longer. - Other correlates: - More tranches, larger amount, non-USD denomination, and term loans are associated with longer maturities. - Leveraged and highly leveraged borrowers have lower maturities compared to investment grade borrowers. - Deals with a guarantor have longer maturities. - Syndicate concentration is associated with significantly longer maturities. --- ### 4 Extensions and Robustness - Matching (Section 4.1): - Nearest-neighbor matching with exact matching on loan type and covariates used in baseline model confirms main results. - After matching, average treatment effect (ATE): - Loans with MDBs’ participation are priced at a higher all-in spread by 33 bps (close to baseline Table 3, column 5). - Difference driven exclusively by higher spread; fees not statistically different. - MDB involvement associated with longer maturities and lower loan volume. - Table 9: treating Risky as treatment and splitting sample by MDB participation: - Risky loans pay a premium of 204 bps on the all-in spread when syndicate is only commercial banks. - Premium decreases to 130 bps when MDBs participate in the syndicate. - Robustness (Section 4.2): - Sensitivity to sample composition: - Dropping borrowers headquartered in China; in China, India and Mexico; and in the 74 countries with less than 50 deals over 1994-2015 does not change main findings on MDBs’ role in pricing and de-risking. - Separate analysis for credit lines and term loans: - MDBs more often involved in term loans—63 percent of deals with MDBs are term loans. - For term loans, MDB involvement significantly lowers borrowing costs for risky borrowers; overall MDBs are not associated with higher borrowing costs on average for term loans. - For credit facilities, reduction of borrowing costs for risky borrowers when MDBs are involved is smaller and less robust. - Alternative clustering of standard errors at country level rather than country-year level: findings remain statistically significant; changes in standard errors are limited. --- ### 5 Conclusions — key findings and mechanisms - Key empirical findings: - MDBs’ participation is associated with higher borrowing costs on average, consistent with MDBs’ greater willingness to finance high-risk projects that private sector may not fund. - MDBs play an important de-risking role, reducing spreads significantly for riskier borrowers and for borrowers in high risk countries. - MDB participation is associated with longer loan maturities and smaller loan size, implying MDBs’ greater capacity to lend at longer tenure and a caveat about the scope for direct mobilization of private resources. - Mechanisms highlighted: - Risk mitigation via better information and monitoring. - Extension of preferred creditor status and informational advantages that can crowd in private investment to developing countries and emerging markets. *Source: wp18263 - 2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts.* ### 2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts . . . . . . . . . . . . . . . 6 ### 2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts ### Data and sample - Sample period: 1994-2014. - Total syndicated loans after cleaning: 16,847 syndicated loans. - Borrowers in sample: 7,589 borrowers headquartered in 107 emerging and developing countries. - Pricing subsample: at most 7,571 deals (and 3,703 borrowers). - Sample dominated by large emerging markets (China, Brazil, India, Mexico, Indonesia, and Turkey); borrowers from low-income and lower middle-income countries represent more than 30 percent of the sample. - Loans are cross-border syndicated loans (borrower and lenders from different countries); sovereign loans excluded. - Loan size measured in 2011 constant USD and excludes domestic-bank-financed portion. ### Loan pricing, maturity, and size (key statistics) - Baseline pricing measure: all-in interest rate spread (contract spread over LIBOR plus any annual fee and any upfront fee). - Average all-in interest rate spread: 351 bps. - Interquantile range of all-in spread: 180 to 475 bps. - Median loan maturity: 3-year maturity. - Share of loans with maturity of one year or shorter: 27 percent. - Share of loans longer than 10 years: 10 percent. - Median loan size: USD 65 million. - Lower quartile loan size: USD 21 million. - Upper quartile loan size: USD 170 million. ### Forms of MDB participation and prevalence - MDB participation takes two forms: A/B loans and parallel loans. - A/B loans: MDB is lender of record (A Loan) and invites external participants for the B Loan. - Parallel loans: MDB and external sources each conclude separate loan agreements with borrower on MDB-administered project. - Share of deals with at least one MDB in the syndicate: about 10 percent (1,694 deals). - Share of MDB-involved deals that are A/B loans: 63 percent. - Main MDBs in sample: European Bank for Reconstruction and Development and International Finance Corporation together account for 56 percent of MDB-involved deals; other key players include the European Investment Bank, the International Bank for Reconstruction and Development, the African Development Bank, the Inter-American Development Bank, and the Asian Development Bank. ### Stylized comparisons: loans with vs. without MDB participation - Average all-in spread premium for loans with MDB participation vs. private-only syndicates: 96 bps more. - Premium reflects an almost equal difference in interest rate spread and in fees. - Average maturity difference: loans with MDB participation have 32 months longer maturity. - Average size difference: deals with MDB participation are smaller by about USD 28 million. - Industry concentration: MDB participation is widespread across industries with a concentration in agriculture and lower presence in manufacturing and natural resources. - Geographic concentration: MDB participation is more common in lending to low- and lower middle-income countries than to emerging markets. - Interpretation: higher spreads and longer maturities are consistent with MDBs’ capacity to lend at longer tenure and with a higher propensity to finance riskier projects (especially infrastructure) that private sector may avoid. ### De-risking evidence and interpretation - Presence of an MDB in a syndicate is associated with a 37 percent reduction in the premium paid by risky borrowers, suggesting a de-risking role (lowering borrowing costs for risky firms). - Possible channels for de-risking: - Informational advantages and strong monitoring capacity of MDBs. - Extension of MDBs’ de facto preferred creditor status to syndicate participants (noted especially in A/B arrangements). - Results control for deal characteristics and absorb time-varying unobserved heterogeneity at industry and country level, including country×industry fixed effects; results robust to sub-samples and matching techniques. ### Macro trends in cross-border syndicated lending and MDB role - Cross-border syndicated loan flows show a cyclical trend: increasing in early 1990s, decline in early 2000s, rapid surge until the global financial crisis, then slight decline. - MDB participation follows a similar pattern and becomes relatively more important during downward phases of the cycle (early 2000s and post-global financial crisis), consistent with a counter-cyclical role. - During downward phases, loans with MDB participation amounted to up to 15 percent of all cross-border lending. - In more recent years (within sample), this share declined to below 10 percent. - Regional shift since 2007: increasing importance of cross-border syndicated lending to low-income countries, especially in South Asia and Sub-Saharan Africa; lending to low-income countries has a strong infrastructure financing component but remains concentrated in a few recipients. *Source: wp18263 - 2.1 MDBs’ Participation in Syndicated Loans: Stylized Facts.* ### 3.2 Loan Pricing ### 3.2 Loan Pricing ### Empirical specification and explanatory power - Table 3 estimates equation 1 with the all-in spread as dependent variable. Column 1 to 5 incrementally add fixed effects up to the preferred specification with country×year and industry×year fixed effects; column 6 includes country×industry fixed effects. - The comparison of the R2 across specifications: - R2 increases from 0.38 to 0.51 between column 1 and 2, indicating global shocks (year fixed effects) play a key role. - Time-varying country and industry fixed effects raise the R2 to 0.66 (column 5), suggesting the model captures two third of the observed variation in loan prices across borrowers. ### MDB participation and pricing - The coefficient on the MDBs’ participation dummy is always positive and statistically significant across specifications, ranging from 82 (column 2, with country and year fixed effects) to 45 (column 5, with country×year and industry×year fixed effects). - Preferred/conservative specification (column 5) implies: - Loans with MDBs’ participation are more expensive by 45 bps or 13 percent (relative to the average all-in spread of 351 bps). - Interpretation: MDBs may self-select into loans with higher risk and higher spreads that private sector would not finance. ### Deal characteristics and pricing - Maturity: an additional year of maturity is associated with a 9 bps increase in the all-in spread (column 5). - Loan size and complexity: - Smaller loans and those with longer maturity are associated with higher prices. - A higher number of tranches is associated with higher prices. - Borrower credit risk: - Highly leveraged borrowers pay on average 365 bps more than investment grade borrowers (column 5). - Leveraged borrowers pay a premium of 115 bps (column 5). - Currency: Loans in Euro and in other currencies have a discount compared with loans in USD. - Loan type: Term loans cost 37 bps more than a credit facility (column 5). - Borrower sector: Public sector companies and government pay 55 bps less than private sector ones. - Guarantor: Deals with a guarantor do not show statistical difference in price from loans without guarantor. - Syndicate concentration (HHI): One standard deviation in the HHI is associated with a discount of 26 bps (column 5). ### Spread versus fees - Results hold when using spread and fees separately (Table 4). - Premium due to MDBs’ presence is almost equally split: - Higher fees: 19 bps. - Higher loan spreads: 25 bps. - Maturity, term loans, and syndicate concentration have larger effects on fees than on spreads. --- ### 3.3 De-Risking ### Hypothesis and identification - MDBs may participate when private market cannot provide funding due to high borrower risk; they also may reduce borrowing costs through de-risking measures (informational advantages, better monitoring, de facto senior creditor status). - Two proxies for borrower creditworthiness used in interactions with the MDB participation dummy: - Risky: dummy = 1 for leveraged or highly leveraged deals. - High country risk: dummy = 1 for borrowers in countries in the bottom half of the Institutional Investor country credit rating distribution. ### Descriptive and regression evidence - Descriptive: For deals without MDBs, strong association between borrower riskiness and all-in spread; with MDBs present, risky borrowers obtain loans priced similar to less risky borrowers. - Table 5 regression results (country×year and industry×year fixed effects): - MDBs’ participation reduces the premium associated with leveraged and highly leveraged borrowers by about a third (column 1). - De-risking effect remains when considering spread and fees separately; effect larger and more precisely estimated for spreads (columns 3 and 5). - Presence of an MDB is associated with significantly lower borrowing costs (about 41 bps) for companies headquartered in riskier countries, controlling for deal characteristics including borrower creditworthiness. --- ### 3.4 Infrastructure and Public Sector Lending ### Infrastructure projects - Interaction of MDB participation with infrastructure project indicator yields a positive and significant coefficient: - Infrastructure loans with MDBs’ participation are about 66 bps more expensive than similar loans financed entirely by commercial banks (Table 6, column 1). - This premium is mostly driven by changes in spread rather than fees (columns 3 and 5). - Interpretation: MDBs tend to finance infrastructure projects with higher risks compared to similar projects financed by commercial banks alone. ### Public versus private sector borrowers - MDB participation is associated with significantly lower borrowing costs for public sector firms. - Economically sizable effect: presence of an MDB in the syndicate almost doubles the reduction in the all-in spread for public sector borrowers compared to private sector ones (Table 6, column 2). - Effect is almost equally large across spread and fees, though the fee estimate is not statistically significant (column 6). --- ### 3.5 Other Loan Terms ### Loan size - MDB participation is associated with lower loan size. - Column 2 estimate: deals with MDBs are on average USD 70 million smaller than loans granted only by commercial banks—almost 40 percent smaller than the average loan of about USD 180 million. - Other findings: - Highly leveraged borrowers obtain smaller loans. - Longer and more complex loans have larger loan size. - Term loans are generally smaller than credit facilities by around USD 35 million. - Higher lender concentration is associated with smaller deals. ### Maturity - MDB participation is associated with longer maturities: - Difference ranges from about 25 months to 27 months across specifications (columns 4-6). - Baseline (column 5): loans with an MDB are on average 27 months longer. - Other correlates: - More tranches, larger amount, non-USD denomination, and term loans are associated with longer maturities. - Leveraged and highly leveraged borrowers have lower maturities compared to investment grade borrowers. - Deals with a guarantor have longer maturities. - Syndicate concentration is associated with significantly longer maturities. --- ### 4 Extensions and Robustness ### Matching (Section 4.1) - Nearest-neighbor matching with exact matching on loan type and covariates used in baseline model confirms main results. - After matching, average treatment effect (ATE): - Loans with MDBs’ participation are priced at a higher all-in spread by 33 bps (close to baseline Table 3, column 5). - Difference driven exclusively by higher spread; fees not statistically different. - MDB involvement associated with longer maturities and lower loan volume. - Table 9: treating Risky as treatment and splitting sample by MDB participation: - Risky loans pay a premium of 204 bps on the all-in spread when syndicate is only commercial banks. - Premium decreases to 130 bps when MDBs participate in the syndicate. ### Robustness (Section 4.2) - Sensitivity to sample composition: - Dropping borrowers headquartered in China; in China, India and Mexico; and in the 74 countries with less than 50 deals over 1994-2015 does not change main findings on MDBs’ role in pricing and de-risking. - Separate analysis for credit lines and term loans: - MDBs more often involved in term loans—63 percent of deals with MDBs are term loans. - For term loans, MDB involvement significantly lowers borrowing costs for risky borrowers; overall MDBs are not associated with higher borrowing costs on average for term loans. - For credit facilities, reduction of borrowing costs for risky borrowers when MDBs are involved is smaller and less robust. - Alternative clustering of standard errors at country level rather than country-year level: findings remain statistically significant; changes in standard errors are limited. --- ### 5 Conclusions - Key empirical findings: - MDBs’ participation is associated with higher borrowing costs on average, consistent with MDBs’ greater willingness to finance high-risk projects that private sector may not fund. - MDBs play an important de-risking role, reducing spreads significantly for riskier borrowers and for borrowers in high risk countries. - MDB participation is associated with longer loan maturities and smaller loan size, implying MDBs’ greater capacity to lend at longer tenure and a caveat about the scope for direct mobilization of private resources. - Mechanisms highlighted: - Risk mitigation via better information and monitoring. - Extension of preferred creditor status and informational advantages that can crowd in private investment to developing countries and emerging markets. *Source: wp18263 - 3.2 Loan Pricing* ### References ### wp18263 - References (Tables & Figures Appendix Summary) ### Key empirical findings on MDB participation and loan pricing - Baseline OLS estimates (Table 3): MDB participation associated with higher all-in spreads: - MDB coefficients: 64.3813***; 82.6265***; 63.0023***; 47.0927***; 45.4043***; 60.4875*** (various specifications). - Decomposition into spread and fees (Table 4): - MDB coefficients on Spread: 24.6882***; 30.4665***. - MDB coefficients on Fees: 19.2753**; 29.0835***. - Matched estimator (Table 8) shows a smaller positive effect: - Matched all-in spread: 32.639**. - Unmatched all-in spread: 101.092***. - Sub-sample and robustness results (Table 10): - Dropping China: MDB = 28.8218**. - Dropping CHN, IND, MEX: MDB = 52.0012***. - Dropping small countries: MDB = 33.3147**. - Alternative clustering (Table A3) preserves MDB positive coefficients: 64.3813***; 82.6265***; 63.0023***; 47.0927***; 45.4043***; 60.4875***. ### De-risking role of MDBs (interaction effects) - Interaction of MDB with loan risk (Table 5): - MDB: 63.2427*** and 66.7475*** (columns 1-2). - MDB x Risky: -59.5944*** and -34.5200*** (columns 1-2). - Risky coefficient: 156.8479***; 153.4012*** (columns 1-2). - Interpretation (from coefficients): MDB participation is associated with higher spreads on average but substantially reduces spreads on Risky deals (negative MDB x Risky). - MDB interaction with country risk (Table 5): - MDB x High country risk: -41.2264*; -23.0643** (columns with spread and fees). - Nearest-neighbor matching on De-Risking (Table 9): - Matched all-in spread when MDB=0: 204.300**. - Matched all-in spread when MDB=1: 130.163***. - Unmatched all-in spread when MDB=0: 237.780***. - Unmatched all-in spread when MDB=1: 98.548***. - # treated (Risky): 2338; # controls: 4196. ### Loan terms, size, maturity, and syndicate structure - Descriptive statistics (Table 1, sample 16,847 deals to 107 countries, 1994-2015): - All-in spread: Obs. 7,038; Mean 351.52; S.D. 322.08; Min 3; p25 75; p50 180; p75 286.64; Max 1,200. - Spread: Obs. 7,345; Mean 168.91; S.D. 116.24. - Fees: Obs. 7,228; Mean 179.87. - Deal value (USD million): Obs. 16,847; Mean 177.03; S.D. 438.64. - Log of deal value: Obs. 16,847; Mean 7.87; S.D. 1.61. - Maturity (months): Obs. 14,915; Mean 52.64; S.D. 61.12. - MDB dummy: Obs. 16,847; Mean 0.10; S.D. 0.30. - Risky dummy: Obs. 16,847; Mean 0.25; S.D. 0.43. - Syndicate concentration (HHI): Obs. 16,847; Mean 0.45; S.D. 0.38. - Average differences by MDB participation (Table 2, 1994-2015): - All-in spread: Commercial banks only mean 345.23 (Obs. 6,600) vs with MDBs’ participation mean 446.32 (Obs. 4,371); Difference 101.09***. - Spread: 165.77 vs 217.38; Difference 51.61***. - Fees: 176.99 vs 223.08; Difference 46.09***. - Deal value (USD million): 179.36 vs 156.18; Difference -23.18***. - Maturity (months): 50.63 vs 81.62; Difference 30.99***. - Loan terms regressions (Table 7): - MDB associated with smaller deal size: MDB = -60.3707**; -69.5282**; -60.8277* (columns 1-3). - MDB associated with longer maturities: MDB = 25.4464***; 26.8113***; 26.2506*** (columns 4-6). - Additional loan-terms robustness/sub-samples (Table A4): - MDB coefficients on deal value: -62.4624**; -40.5714; -53.8281**. - MDB coefficients on maturity: 14.5607***; 15.3793***; 15.1415***. ### Infrastructure, public sector lending, and MDBs - Interaction with infrastructure and public sector (Table 6): - MDB baseline: 32.6194**; 55.2291*** (all-in spread columns). - MDB x Infrastructure: 65.6771** (All-in Spread); 39.6436*** (Spread). - MDB x Public: -52.9012* (All-in Spread); -26.9046** (Spread). - Public coefficient: -55.7821***; -52.6893*** (all-in spread columns). - Sector composition (Table A2): - Infrastructure deals: # deals 3,791; % of total 0.23; % of which, with MDB 4,360 (?) — (table presents counts and percentages; preserve values as in source). ### Syndicate characteristics and other controls with consistent effects - Number of tranches associated with higher all-in spreads (Table 3): - Coefficients: 21.2247***; 15.1710**; 11.8910**; 13.9997***; 13.8604***; 11.0556*. - Maturity per month increases spreads modestly (Table 3): coefficients 0.4245*; 0.7051***; 0.7194***; 0.7929***; 0.7700***; 0.6750***. - Leveraged and Highly leveraged dummy effects (Table 3, Table 4): - Leveraged: 171.4569***; 159.1101***; 130.6725***; 119.2904***; 115.4159***; 128.8962***. - Highly leveraged: 426.2991***; 465.8138***; 388.7043***; 368.8614***; 364.6687***; 381.8506***. - Public borrower dummy associated with lower spreads (Table 3, Table 4): - Public: -37.1885**; -49.718***; -55.140***; -61.913***; -55.473***; -54.526***. - Currency effects: - Euro loan tends to have negative coefficients on spreads (e.g., -32.9701; -43.8180**; -49.667*** across specifications). - Other currency coefficients often negative and significant (e.g., -43.774***; -47.092***). ### Matching and nearest-neighbor estimates (Table 8 summary) - Matched sample effects (nearest neighbor): - All-in spread: Matched 32.639** (standard error 16.516). - Spread: Matched 30.274*** (standard error 7.746). - Fees: Matched 9.660 (not significant). - Deal size: Matched -83.297***. - Maturity: Matched 26.342***. - Matched vs Unmatched contrast: - Unmatched all-in spread: 101.092***. - # treated (MDB): 424, 433, 438, 948, 948 (depending on specification). - # controls vary across specifications (e.g., 653, 468, 1767, 7071, 3967). ### Figures and descriptive trends (as presented) - Density comparisons of all-in spreads (Figures 1 and 3): - Figure 1: density of all-in spread (bps) for cross-border syndicated loans to developing countries; sample 7,038 deals to 106 countries; separates deals With MDBs’ participation vs Without MDBs. - Figure 3: shows densities for Investment grade vs Leverage and highly leveraged, separately for deals without MDBs’ participation and deals with MDBs’ participation; sample 7,038 deals to 106 countries. - Time series of cross-border syndicated lending to developing countries in USD (Figure 2): - Value in constant 2011 USD (billion), sample of 16,847 deals to 106 countries; separates Private banks loans vs Loans with MDBs participation. *Source: wp18263 - References (figures and tables) PDF.* --- _Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18263.pdf_