On the Role of the Credit Rating Agencies in the Euro Zone Crisis

Abstract

We examine the determinants of credit ratings for 18 Eurozone countries over the period 2002–2013. Sovereign credit ratings are decomposed into high and low ratings, the high rated being AA− and above, and the low rated being A+ and below. We consider a set of macroeconomic and risk variables as their determinants. First, we find greater explanatory power for the former sample (high rated). Second, the results reveal an asymmetric response of cumulated current account for high and low ratings. Third, we provide evidence that the fiscal and the external sector are significant after 2009 only for the low rated economies. Focusing on Greece we provide evidence that the government debt and cumulative current account played a significant role in the downgrade of Greek bonds.

Appendix

Table A1

Authors	Dependent variables	Explanatory variables	Methodology/data	Important results
Cantor and Packer (1996)	Moody’s rating S&P rating	Per capita income GDP growth Inflation Fiscal balance External balance External debt Economic development Default history	49 countries in 29 September 2001. Moodys and S&P and its average. Cross-section OLS	Per capita income, inflation, external debt, economic development and Default History explain more than 90% of the variation of credit rating for Moodys, S&P and its average
Haque et al. (1998)	Institutional Investor Euro money Economic Intelligence unit	Economic variables: Terms of trade Export growth Current account/GDP Reserves/Imports External debt/GDP Real exchange rate Growth Inflation Political variables Coups Assassination General strikes Guerilla warfare Government crises Purges	Cross section OLS	Credit rating appears to be determined mainly through the analysis of economic variables. Political variables do not add any additional explanatory power
Eliasson (2002)	S&P rating	Per capita income GDP growth Inflation Fiscal balance External balance External debt Economic development Default history Short-term currency Debt to foreign reserves Export growth Interest rate spreads	38 emerging countries from 1990 to 1999. Static and dynamic panel model	Dynamc model is more robust than the static one. Using static panel data model both spreads and short-term debt to reserves are significant variables. Current account entered in both models with an unexpected sign
Bissoondoyal Bheenick (2005)	S&P rating Moody’s rating	Per capita income Inflation Govt financial balance/GDP Govt debt/GDP Real exchange rate Foreign reserve Net Exports/GDP Unemployment rate Unit labor cost Current account/GDP Foreign debt/GDP	95 countries from December 1995 to December 1999 Ordered Response Model. First using rating from 1 to 9 and then from 1 to 21. Estimated first full sample, second for high rated countries and third for low rated countries	Economic variables do not play an important role for the high rated sample of countries. GNP per capita and inflation are the most significant factors for the full sample. Moreover, current account balance and the level of foreign reserves do play an important role for low rated countries
Mellios and Paget-Blanc (2006)	S&P rating Fitch rating Moody’s rating	Per capita income GDP growth Inflation Economic development Current account Default history Real exchange rate Foreign debt/GDP Ratio debt/GDP Ratio reserves/imports Ratio investment/GDP2Corruption index Regulatory quality	86 countries in December 31 2003. Cross section OLS Ordered Logistic Model	OLS suggest that 11 explanatory variables are statistically significant. In contrast ordered logistic model suggest only nine. logistic model behaves better than the OLS model
Afonso (2003)	Moody’s Rati ng S&P Rating	Per capita GDP Inflation rate GDP real growth rate Development indicator Default indicator External debt-exports ratio Government deficit as a percentage of GDP	Cross-section OLS using both a linear and a logistic transformation of the data. 81 countries in June 2001	Logistic transformation turned out to be better for the overall sample, especially for the countries placed on the top end of the rating scale. GDP per capita, external debt, economic development, default history, real growth rate and the inflation rate explained a big part of the variability of credit ratings
Rowland (2004)	Moody’s rating S&P rating EMBI Global composite Institutional Investor’s creditworthiness Index	GDP per capita Real GDP growth rate Fiscal balance as a percentage of GDP Current account balance as a percentage of GDP Debt-to-GDP ratio Debt ratio International reserves as a percentage of GDP Debt-service-to-GDP ratio Openness Inflation rate Default history	49 countries at the end of July 2003. Moodys and S&P OLS regression for sovereign credit ratings, sovereign spreads and creditworthiness	GDP per capita is a significant explanatory variable in all the regressions. Regression on the determinants of the creditworthiness index has by far the lowest adjusted R value
Valle and Marin (2005)	Moody’s rating Fitch rating S&P rating	GDP per capita GDP growth Increase of the CPI Fiscal balance/GDP Balance of payments on current account/GDP Internal debt of the state/GDP Liquidity ratio Industrialization	80 countries dated 28 of March 2003. OLS regression using first 9 explanatory variables and the n 4 and 5	The model with 4 explanatory variables has as much power as the others. GDP per capita, GDP growth and inflation found to be statistically significant and with the expected signs
Bautler and Fauver (2006)	Institutional investor Moody’s rating S&P rating Ten year sovereign bond yields	GDP per capita Inflation Underdevelopment Index Default dummy Voice of the people Political stability Government effectiveness Regulatory quality Rule of law Corruption control Legal environment composite Emerging market dummy Foreign debt/GDP Common law dummy	86 countries in March 2004 OLS for th efull sample 2SLS using as instruments for legal environment the ethnolinguistic fractialiazation and French civil law origin. Differentiation acrross low and high debt countries	Using OLS legal environment found to be statistically significant and its marginal effect in sovereign credit rating is much stronger than macroeconomic variables. Using 2SLS the effect of legal environment on credit rating is smaller than OLS estimates indicate, although it is still quite large. Sovereign credit rating are more sensitive in legal environment in low-debt countries than high-debt
Archer et al. (2007)	S&P rating Moody’s rating Fitch rating	Political factors: Presidential ideology Executive party tenure Undivided government Election cycles Honeymoon periods Economic Variables: Total external debt Inflation Gdpper capita Current account balance Default history Natural resources	50 developing countries from 1987 to 2003. Panel-corrected standard errors estimation using both annual bond ratings and two year moving average	All political variables, except from executive party tenure, are found to be statistically insignificant. The measure with the biggest impact is history of bond default in the previous five years. Inflation, Gdp growth rates and trade are highly accounted for the three rating agencies
Afonso et al. (2007)	S&P rating Moody’s rating Fitch rating	Per capita income Real GDP growth Inflation Unemployment Government debt Fiscal balance Government effectiveness External debt Foreign reserves Current account balance Default history	130 countries from 1995 to 2005. Linear panel estimation using pool OLS, fixed effects and random effects. Differentiation across sub periods, 1996–2000 a nd 2001–2005. Differentiation across rating levels, BBB+ and above. Ordered probit estimation and random effects ordered probit estimation fot the full sample	Per capita GDP, GDP real growth rate, government debt, government effectiveness, external debt and external reserves relevant for the determination of the sovereign credit ratings. For the low rating levels, external debt and external reserves are more relevant Inflation plays a bigger role for high rating levels. Moreover, after the Asian crisis, it see ms there was a decline in the relevance of the current account variable in the specifications for Moody’s and S&P
Afonso et al. (2011)	Fitch rating Moody’s rating S&P rating	Per capita income Real GDP growth Unemployment Inflation Government debt Fiscal balance Government effectiveness External debt Foreign reserves Current account balance Default history European Union dummies’ Regional dummies	130 countries from 1995 to 2005. Linear panel random effects estimation. Ordered probit random effects estimation	Per capita GDP, real GDP growth, government debt, and government deficit have a short-run impact on a country’s credit rating. Government effectiveness, external debt, foreign reserves, and sovereign default dummies have only a long run impact
Zheng (2012)	S&P rating Dagong rating	GDP per capita Real GDP growth Inflation Fiscal balance External balance External debt Internal debt Economic development Default history	43 countries in 2011. Linear regression using Dagong, S&P their average and their difference in both local and domestic currency ratings as dependent variable	Agencies use similar economic risk indicators. Inflation, external balance, and the dummies for economic development and default history come out statistically significant in both agencies’ ratings. But Dagong assigns different weights to these indicators
Bozic and Magazino (2013)	Moody’s rating Fitch rating S&P rating	GNI growth Per capita GNI Current account balance Inflation Unemployment Fiscal balance Government debt Real Interest Rate Reserves Default history EMU membership Fiscal balance squared Government debt squared	139 countries in the period 1975–2010. Unbalanced Panel using pooled OLS, fixed effects, random effects and panel corrected standard errors. Differentiation across sub-periods 1975–1996 and 1997–2010 and on the development level	Per capita GNI, inflation, unemployment, fiscal balance, government debt and default history are statistically significant in almost all regressions and for all rating agencies. EMU membership increases rating and both fiscal balance and government debt square are strongly significant
Garcia et al. (2014)	Moody’s rating Fitch rating S&P rating	Per capita income GDP growth Inflation Fiscal balance External balance External debt Economic development Previous payment behaviour Control of corruption Government effectiveness Political stability and absence of violence Regulatory quality Rule of law Voice and accountability	82 countries 2004–2011 OLS. First with 14 explanatory variables and then only with the three statistically significant of the First regression	External balance, economic development and regulatory quality are statistically significant
Boumparis et al. (2015)	Moody’s rating Fitch rating S&P rating Average rating	GDP per capita GDP growth rate Government debt Inflation rate Unemployment rate Cumulated current account External balance Reserves Regulatory quality	18 Euro zone countries from 2002 to 2013. Pooled OLS, fixed effects, random effects using cross sectional averages as additional explanatory variables	All variables except for cumulated current account are statistically significant government debt and cumulative current account exert a stronger positive impact on credit ratings post-2008. No remaining cross sectional dependence