63889 POVERTY THE WORLD BANK REDUCTION AND ECONOMIC MANAGEMENT NETWORK (PREM) Economic Premise August 2011 • Number 63 JUNE 2011 • Number 60 JUN 010 • Numbe 18 Shadow Sovereign Ratings How Complementary Are Prudential Regulation and Monetary Policy? Otaviano Canuto, Sanket Mohapatra, and Dilip Ratha Otaviano Canuto Could either monetaryapolicy or financial prudential regulation be relied on individually to mitigate asset price cycles or Sovereign ratings are necessary condition for countries to fully access international capital. Even if the sovereign government both issuing bonds, the monetary policy and acts as a “ceiling” for could then be considered influence its their effects? Ifis not ways are effective, sovereign rating often prudential regulationthe private sector and can “substitutes,” in international capital market access. However, 58 developing reduction in still not rated by Standard & Poor’s, Moody’s, the sense that the individual use of either instrument leads to a countries are the volatility of both corresponding targets. This and Fitch, the three in favor of complementarity—rather than substitution—in the use of monetary “shadow” sovereign note, however, arguesinternational credit rating agencies. This premise presents an exercise to predictand macroprudential ratings to combined (articulate) countries monetary and credit spectrum if they were rated. tends to be popular policies: theestimate where unrateduse of bothwould lie on themacroprudential policies and rulesContrary tomore effective perception, unrated countries are of necessarily at the bottom of the rating spectrum. than a standalone implementationnoteither. , Monetary Policy Asset Prices, and Finan- Introduction “Great (Borensztein, Cowan, and economies, with There- countryModeration” in developedValenzuela 2007). relatively cial Stability low the country and small as a benchmark for the interna- fore,inflation ratesrating acts output fluctuations from the mid- Sovereign ratings not only are important for attracting 1980s onward, seems to vindicate this path. tional capital market activities of the private sector. Asset price cycles had been a concern for many years prior to private capital flows, but also act as widely available and in- As is now known, this world of presumed stable are not However, as of mid-2011, 58 developing countriesmonetary the recent global financial crisis, but were seen as a separate is- ternationally comparable indicators of a country’s fiscal per- and financial conditions was severely and Fitch, the three in- rated by Standard & Poor’s, Moody’s, shaken by the recent glob- sue that was not a monetary policy concern. Even when the formance. A country’s sovereign rating provides a basis for al financial crisis. With the benefit 36 countries have had ternational rating agencies. Another of hindsight, it is easy to frequent appearance of asset price bubbles started to be ac- international investors and bondholders to assess the risks draw lessons. Asset price booms and busts Countries in the the same assigned rating since early 2009. were acknowledged knowledged, the belief was—“the Greenspan-Bernanke ap- of a country’s ability to honor its public debt obligations to be both pervasive creditworthiness evaluated stock market first group need theirand harmful: real estate and to improve proach”1—that attempts to detect and prick them at an early (Beers and Cavanaugh 2005; Lehmann 2004; Truglia and booms contributed to excess international financing. Coun- their access to market-based U.S. household debt and to fragile stage would be impossible and potentially harmful. If neces- Cailleteau 2006). Assessments of sovereign creditworthi- asset liability structures, the interconnectedness sovereign tries in the latter group need to have their current of financial ness are also up after the bubble types of resource flows, in- sary, mopping important for other burst would be safer, using firms’ balance to determine whether that rating is justified rating assessed sheets, and the danger of too-big-to-fail institu- interest rate cuts to help economic recovery.2 cluding official aid (for example, performance-based aid al- tions. The macroeconomic fundamentals or price bust by current rapid global transmission of an asset whether Low, stable inflation is a necessary and sufficient condition location by the U.S. Millennium Challenge Account) and pushed in world economy to the edge of quasi-collapse (Ca- changes the the country’s policy or institutional variables for stable growth with moderate unemployment. This condi- concessional loans provided by multilateral and bilateral nuto suggest might2009). an upgrade (or downgrade, as appropriate) of tion could be pursued, among other ways, through an inflation donors. existing it lax the But was rating.monetary policy that led to the creation of Even when the sovereign government is clear communica- targeting framework, using interest rates and not issuing bonds, This premise presents an exercise instability? Some, such these bubbles and then to financial to predict “shadow” sov- as the sovereign rating predefined a “ceiling” for the foreign sin- tion rules to achieve aoften acts as inflation objective, as thecur- Svensson (2010), say no. For unrated countries would was ereign ratings to estimate wherethem, the financial crisis lie rency rating of bonds authorities. Stable banks located in the gle focus for monetary issued by firms and inflation would also caused by factors other than monetary policy; monetary policy on the credit spectrum if they were rated.1 result in low-risk premia, which combined with competition in and financial stability policy are distinct–it was the latter that financial markets would help achieve financial stability. The failed.3 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise A Brief History of Sovereign rating process can be complex and not readily available in many Credit Ratings countries. The institutional and legal environment that governs property rights and the sale of securities may be absent or weak, Sovereign credit ratings have existed for nearly a century. Two which prompts reluctance on the part of politicians to be pub- of the major rating agencies—Standard & Poor’s and Moody’s— licly judged by the rating analysts. Some countries find it dis- started rating sovereign Yankee bonds in the early 20th centu- couraging to request a rating, pay a fee for the rating, and then ry. By 1929, 21 countries were rated by Poor’s Publishing, the have no command over the final outcome. Basel capital ade- predecessor to Standard & Poor’s, including several of today’s quacy regulations that assign a lower risk weight (100 percent) emerging markets, such as Argentina, Colombia, and Uruguay to unrated entities than to those rated below BB− (150 percent) (Bhatia 2002). Moody’s started rating debt instruments in may also discourage borrowing entities from being rated. 1919, and within the next decade, it had rated bonds issued by about 50 governments (Cantor and Packer 1996). However, A Predictive Model for Sovereign Ratings demand for ratings declined during the Great Depression, and most ratings were suspended following World War II. Rating Many researchers have found that ratings by the major agencies activity for sovereigns resumed in the 1970s but at a signifi- are largely explained by a handful of macroeconomic variables cantly slower pace until the 1980s. In 1980, eight high-income (Cantor and Packer 1996; Canuto, dos Santos and de Sá Porto countries were rated by at least one of the three leading rating 2004; Lee 1993; Ratha, De, and Mohapatra 2011; Rowland agencies. By the late 1980s, almost all the high-income Organ- 2005). Ferri, Liu, and Stiglitz (1999) and Mora (2006) used isation for Economic Co-operation and Development coun- similar models to examine whether ratings were procyclical tries had been rated. during the Asian crisis by comparing predicted with actual rat- Sovereign credit ratings for developing countries (as currently ings. Related literature has found that a small set of variables defined by the World Bank) began in the late 1980s after the sov- explains the likelihood of debt distress and defaults (Kraay and ereign debt crises earlier that decade. The number of rated devel- Nehru 2006; Reinhart, Rogoff, and Savastano 2003).2 oping countries increased significantly during the 1990s emerg- The first step in the empirical analysis is to convert the let- ing market phenomena. By April 2011, 135 countries—45 ter long-term foreign currency rating from the three major high-income and 90 developing countries—were rated by at least agencies to a numerical equivalent (Bhatia 2002; Canuto, dos one of the three agencies. Furthermore, sovereign ratings issued Santos, and de Sá Porto 2004). In the scale used for this exer- by different agencies tend to be highly correlated. The bivariate cise (see table 1), 1 denotes the highest rating (corresponding correlation coefficient between the ratings of the three agencies to AAA for Standard & Poor’s and Fitch, Aaa for Moody’s) ranges from 0.97 to 0.99. For most developing countries, the rat- and 21 denotes the lowest rating (or C for all three agencies). ings are usually within one to two notches of one another. Cases of sovereign or selective default are excluded in this re- gression analysis because assigning a specific numeric rating The Case for Predicting Shadow Ratings to such extreme credit events is difficult. Although default or for Unrated Developing Countries selective default appears to be just another step down the road of getting a rating downgrade, assigning a specific value to Most of the unrated countries need capital from international such an event would risk ignoring the degree of distress (for markets. Yet without a credit rating, those countries have dif- example, a temporary liquidity crisis versus a systemic crisis). ficulty accessing international bond markets and resort to The next step is to estimate the numeric equivalent of sov- costly relationship-based borrowing from commercial banks ereign ratings for the rated developing countries as a function or to sales of equity to foreign direct investors. This scenario is of macroeconomic variables, rule of law, debt and interna- especially true for subsovereign entities and private companies tional reserves, and macroeconomic volatility (as identified in for which the sovereign rating acts as a ceiling (Canuto and Liu the literature). A linear regression model of the data is pre- 2010a, 2010b). Without a sovereign rating, such borrowers sented in the following equation: tend to be cut off from international credit markets. Thus, it is Sovereign rating = α + β1(log of GNI per capita) + β2(GDP in the interest of all countries to obtain a credit rating even if growth rate) the sovereign government does not need to borrow. + β3(Debt/Exports) + β4[Reserves/(Imports + Short- Why are so many countries not rated in the first place? Sev- term debt)] eral factors influence a country’s reluctance or inability to get + β5(Growth volatility) + β6 (Inflation) + β7(Rule of rated. Countries are constantly reminded of the risks of cur- law) + error (1) rency and term mismatch associated with market-based foreign currency debt, as well as the possibility of sudden reversal of in- Data for most of the right-hand variables are from the vestor sentiment. The information required for the commercial World Bank’s World Development Indicators database and 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Table 1. Sovereign Ratings: Conversion from Letter to Numeric Scale Standard & Poor’s Fitch Moody’s Numeric grade Investment grade Highest credit quality AAA AAA Aaa 1 Very high credit quality AA+ AA+ Aa1 2 AA AA Aa2 3 AA− AA− Aa3 4 High credit quality A+ A+ A1 5 A A A2 6 A− A− A3 7 Good credit quality BBB+ BBB+ Baa1 8 BBB BBB Baa2 9 BBB− BBB− Baa3 10 Speculative grade Speculative BB+ BB+ Ba1 11 BB BB Ba2 12 BB− BB− Ba3 13 Highly speculative B+ B+ B1 14 B B B2 15 B− B− B3 16 High default risk CCC+ CCC+ Caa1 17 CCC CCC Caa2 18 CCC− CCC− Caa3 19 Very high default risk CC CC Ca 20 C C C 21 Sources: Standard & Poor’s, Moody’s Investors Service, and Fitch Ratings. the International Monetary Fund’s World Economic Outlook Figure 1. Distribution of Predicted Ratings database, which are now publicly available. Data on short- and long-term claims are collected from the Bank of Interna- 20 tional Settlements. The rule of law variable is taken from a widely used dataset produced and updated by Kaufmann, Kraay, and Mastruzzi (2009). The signs of the explanatory 15 variables are in the expected direction and are significant at number of countries the 10 percent level or better (see Ratha, De, and Mohapatra 2011). All the variables together explain about 80 percent of 10 the variation in ratings for the regression sample.3 Shadow Ratings for Unrated Developing 5 Countries This exercise uses the benchmark model to predict ratings for 0 investment BB B CCC or lower the unrated developing countries. The results are presented grade in the annex. Strikingly, the predicted ratings for the unrated sovereign rating countries do not all lie at the bottom end of the rating spec- Source: Authors’ calculations. trum but are spread over a wide range (figure 1). Note: The distribution is based on the lowest predicted rating. 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Table 2. Comparison of Actual and Predicted Ratings for Ratings Issued after January 2007 Country Sovereign rating Date established Shadow rating (April 2011) Angola B+ May 2010 B to B+ Bangladesh BB− to BB April 2010 B+ Belarus B March 2011 B+ to BB− Gabon BB− November 2007 BB to BB+ Libya BB March 2011 BB+ or lower Rwanda B August 2010 B to B+ Zambia B+ March 2011 BB Sources: Authors’ calculations, Fitch, Moody’s, and Standard and Poor’s. Of 47 unrated countries from an original 55 unrated coun- ciently to deserve an upgrade (or changed enough to require a tries for which Ratha, De, and Mohapatra (2011) generated downgrade). The shadow ratings also suggest a group of indi- predicted ratings, 7 countries are likely to be investment grade, cators that developing countries can improve to achieve a 10 are likely to be in the BB category, 20 in the B category, and higher sovereign rating. 10 in the CCC or lower category. The countries just below the The international donor community can play a role in investment grade but at or above CCC are comparable to many helping developing countries to obtain ratings. Such policy emerging market countries with regular market access. For ex- interventions have precedents. The United Nations Develop- ample, in our analysis, Swaziland’s shadow rating from Stan- ment Programme partnered with Standard & Poor’s to rate dard & Poor’s ranges from B+ to BB, which puts the country in eight African countries during 2003–06 (Standard & Poor’s a similar bracket as Indonesia. Several other unrated develop- 2006), several of which have since accessed international cap- ing countries (for example, Algeria, Bhutan, Djibouti, Equato- ital markets to raise financing at a lower cost than the domes- rial Guinea, Maldives, and the Syrian Arab Republic) have tic borrowing cost. shadow ratings in the B category or above. Knowing the shadow ratings of unrated countries can also While the predicted or shadow rating indicates the likeli- be helpful to bilateral and multilateral donors interested in hood of default on foreign currency debt obligations of the setting up guarantees and other financial structures to reduce sovereign, it is not a predictor of whether the country will be project risks and to mobilize private financing. One such in- successful if it were to issue an international bond. This is par- novative financing instrument that is being discussed is dias- ticularly true for small countries where volatility of economic pora bonds to tap into the considerable wealth of the diaspora growth and government revenue can be too high to render of developing countries (Okonjo-Iweala and Ratha 2011). them unable to access private capital markets.4 These mechanisms can complement existing efforts to im- Table 2 presents the shadow ratings for several countries prove aid effectiveness. that were rated since the estimates of Ratha, De, and Mohapa- tra (2011) in early 2007. The predicted ratings are within one About the Authors notch of the actual rating range for five of the seven countries. The difference between the predicted and actual ratings likely Otaviano Canuto is vice president of the Poverty Reduction and reflects improvement (or deterioration) in macroeconomic Economic Management (PREM) Network of the World Bank. fundamentals during the intervening period. Sanket Mohapatra is an economist in the Migration and Remit- The model-based shadow ratings can provide a benchmark tances Unit of the World Bank. Dilip Ratha is a lead economist in for evaluating unrated countries or rated countries that have the Development Prospects Group and the manager of the Migra- not been rated for some time and might have improved suffi- tion and Remittances Unit of the World Bank. 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Annex: Shadow Ratings for Unrated Countries, April 2011 Country Shadow rating (April 2011) Rated countries in a similar range Algeria BB to BB+ Indonesia, Turkey Bhutan BB− to BB Bangladesh; Venezuela, RB Burundi C or lower Gambia, The; Malawi Central African Republic CCC+ to B− Belize, Zambia Chad CC to B Belize, Ecuador Comoros CCC− to CCC+ Gambia, The; Malawi Congo, Dem. Rep. CCC− to CCC Gambia, The; Malawi Congo, Rep. B+ to BB− Guatemala, Uganda Côte d’Ivoire B− or lower Ecuador, Pakistan Djibouti B+ to BB− Guatemala, Uganda Dominica BB+ to BBB Costa Rica, Croatia Equatorial Guinea B+ to BB Dominican Republic, Paraguay Eritrea CCC− to CCC Gambia, The; Malawi Ethiopia B− to B Jamaica, Mali Guinea C to CCC− Gambia, The; Malawi Guinea-Bissau CCC+ to B Belize, Zambia Guyana B+ to BB− Guatemala, Indonesia Haiti B− to B Mali, Jamaica Iraq B Honduras, Ghana, Burkina Faso Kiribati* A+ China Kyrgyz Republic CCC+ to B− Belize, Zambia Lao PDR B− to B+ Argentina, Belarus Liberia CCC+ to B Belize, Zambia Maldives B+ to BB+ Latvia, Senegal Marshall Islands* B− to B+ Jamaica, Mali Mauritania B− to B Jamaica, Mali Myanmar CCC+ to B− Belize, Zambia Nepal CCC+ Gambia, The; Malawi Niger B− to B+ Argentina, Belarus Samoa BB+ to BBB Costa Rica, Croatia São Tomé and Príncipe CCC or lower Gambia, The; Malawi Sierra Leone CCC+ to B− Belize, Zambia Solomon Islands* B− to B+ Jamaica, Mali St. Kitts and Nevis* BBB+ to A Brazil, Panama St. Lucia BBB− to A− Botswana, Panama St. Vincent and the Grenadines BB+ to BBB Costa Rica, Croatia Sudan CCC− to CCC+ Gambia, The; Malawi Swaziland B+ to BB Dominican Republic, Indonesia Syrian Arab Republic BB− to BB+ Uruguay, Vietnam Tajikistan C to CCC Gambia, The; Malawi Tanzania B+ Albania, Angola, Kenya Togo B− to B+ Argentina, Belarus Tonga* B+ to BB+ Colombia, Indonesia Uzbekistan B to B+ Bolivia, Lebanon Vanuatu BBB− to BBB+ Kazakhstan, Mexico Yemen, Rep. B− to B Jamaica Zimbabwe CC to CCC− Gambia, The; Malawi Source: Updated from Ratha, De, and Mohapatra 2011. Note: Shadow ratings for unrated countries marked with an asterisk (*) are from Ratha, De, and Mohapatra (2011). The model-based ratings should be treated as indicative; they are clearly not a substitute for the broader, deeper analysis and qualitative judgment employed by experienced rating analysts. The predicted ratings range is based on predictions for the benchmark models for Standard & Poor’s, Moody’s, and Fitch. Forecasts of explanatory variables for 2011 (as available in April 2011) were used to predict ratings for 2011. Predicted ratings for rated countries were also generated and are available upon request. 5 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Notes Borensztein, E., K. Cowan, and P. Valenzuela. 2007. “Sovereign Ceilings ‘Lite’? The Impact of Sovereign Ratings on Corporate Ratings in Emerg- ing Market Economies.” Working Paper 07/75, International Monetary 1. The exercise follows an econometric model developed by Fund, Washington, DC. Ratha, De, and Mohapatra (2011) that explains ratings as- Cantor, R., and F. Packer. 1996. “Determinants and Impact of Sovereign signed to developing countries by the three major rating agen- Credit Ratings.” Economic Policy Review 2 (2): 37–53. Canuto, O., and L. Liu. 2010a. “Subnational Finance: Make It Sustainable.”In cies. The shadow ratings are updated to the current year using The Day after Tomorrow: A Handbook on the Future of Economic Policy in International Monetary Fund and World Bank forecasts of the Developing World, ed. O. Canuto and M. Giugale, 219–37. Washing- explanatory variables for 2011. For a previous econometric ton, DC: World Bank. http://www.worldbank.org/prem. exercise using fixed-effects methods, see Canuto, dos Santos, ———. 2010b. “Subnational Debt Finance and the Global Financial Crisis.” Economic Premise 13: 1–7. and de Sá Porto (2004). Canuto, O., P. F. P. dos Santos, and P. C. de Sá Porto. 2004. “Macroeconomics 2. Because most of the unrated countries (for which this exer- and Sovereign Risk Ratings.” Paper presented at a seminar at the School cise predicts ratings) are also low-income countries, this exer- of Economics, Business, and Accounting, University of São Paulo, São Paulo, Brazil, January. cise has some similarities with that of Kraay and Nehru Ferri, G., L. G. Liu, and J. E. Stiglitz. 1999. “Are Credit Ratings Pro-cyclical? (2006). However, this exercise uses a continuous numeric Evidence from East Asian Countries.” Economic Notes 28 (3): 335–55. scale for ratings and excludes cases of default in the regres- Kaufmann, D. K., A. Kraay, and M. Mastruzzi. 2009. “Governance Matters sions, unlike the 0–1 dummy for debt distress used by Kraay VIII: Aggregate and Individual Governance Indicators, 1996–2008.” Policy Research Working Paper 4978, World Bank, Washington, DC. and Nehru. Kraay, A., and V. Nehru. 2006. “When Is External Debt Sustainable?” World 3. Ratha, De, and Mohapatra (2011) test the predictive power Bank Economic Review 20 (3): 341–65. of this model using “within-sample” prediction. This exercise Lee, S. H. 1993. “Are the Credit Ratings Assigned by Bankers Based on the also exploits the high correlation across ratings assigned by Willingness of LDC Borrowers to Repay?” Journal of Development Eco- nomics 40 (2): 349–59. the three agencies to test whether the predicted rating for one Lehmann, A. 2004. “Sovereign Credit Ratings and Private Capital Flows to agency is similar to the actual ratings by other agencies. Low-Income Countries.” African Development Review 16 (2): 252–68. 4. The shadow ratings for some small economies seem unex- Mora, N. 2006. “Sovereign Credit Ratings: Guilty Beyond Reasonable pectedly high. Kiribati’s A+ rating is likely due to extraordi- Doubt?” Journal of Banking and Finance 30 (7): 2041–62. Okonjo-Iweala, N., and D. Ratha. 2011. “A Bond for the Homeland.” Foreign narily high reserves accumulated from earlier phosphate min- Policy, May 24. http://www.foreignpolicy.com/articles/2011/05/24/a_ ing revenues in a Revenue Equalization Reserve Fund. The bond_for_the_homeland. high shadow ratings of Samoa and Vanuatu reflect high levels Ratha, D., P. De, and S. Mohapatra. 2011. “Shadow Sovereign Ratings for Un- of international reserves which in turn depend on the contin- rated Developing Countries.” World Development 39 (3): 295–307. Reinhart, C. M., K. S. Rogoff, and M. A. Savastano. 2003. “Debt Intolerance.” ued availability of official aid. Brookings Papers on Economic Activity 1: 1–75. Rowland, P. 2005. Determinants of Spread, Credit Ratings, and Creditworthi- References ness for Emerging Market Sovereign Debt: A Follow-Up Study Using Pooled Data Analysis. Bogotá: Banco de la República. Beers, D., and M. Cavanaugh. 2005. Sovereign Credit Ratings: A Primer. New Standard & Poor’s. 2006. Sovereign Ratings in Africa. New York: Standard & York: Standard & Poor’s. Poor’s. Bhatia, A. 2002. “Sovereign Credit Ratings Methodology: An Evaluation.” Truglia, V., and P. Cailleteau, P. 2006. “A Guide to Moody’s Sovereign Rat- Working Paper 02/170, International Monetary Fund, Washington, DC. ings.” Special Comment, Moody’s, New York. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. It is produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at http://www.worldbank.org/economicpremise. 6 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise