This blog article summarizes a recent article published in the European Business Law Review.
Reference: Norbert J. Gaillard, “Credible Sovereign Ratings: Beyond Statistics and Regulations”, European Business Law Review, 28(1), 2017.
This blog article has five purposes. Section 1 shows that sovereign ratings accuracy is contingent upon (i) bailout packages that prevent some high-rated governments from defaulting, and (ii) downgrades at a time when the credit position of sovereign issuers is already weakened. Section 2 analyzes why credit rating agencies (CRAs) have been unable to reach a satisfactory trade-off between accuracy and stability. Section 3 investigates how the widespread use of credit ratings among regulators and investors – as well as the business model of CRAs – has led to biased and inflated sovereign ratings. Section 4 discusses the main provisions of Regulation (EU) No.462/2013 of the European Parliament and of the Council of 21 May 2013 devoted to sovereign ratings. Section 5 applies the insights and lessons learned to describe an improved sovereign rating methodology.
- Accuracy of Sovereign Ratings
The accuracy of sovereign ratings can be assessed by examining ratings prior to defaults, computing cumulative default rates, or computing accuracy ratios. These three metrics establish the ability of CRAs to discriminate between good and bad sovereign debtors. However, that performance must be put into perspective.
Unlike most firms in the private sector, a sovereign issuer can stave off default by calling on foreign governments and/or supranational institutions to act as lenders of last resort (LOLRs). For example, had South Korea not been bailed out by the IMF in December 1997, the country would have gone bankrupt. That a sovereign borrower could be rescued enhances its credit rating, as Fitch stated a few months after the Asian crisis.
Yet IMF support packages cannot be taken for granted: deals may be hard to reach even when the distressed country’s default would increase systemic risk levels – as was presumably the case with South Korea. Since 1997, the methodologies of Fitch, Moody’s, and S&P have accounted for this dynamic. However, the focus has been much more on what rating best captures a country’s creditworthiness once within the IMF’s purview than on a government’s ability to secure funding from international institutions and creditors. During the Eurozone debt crisis, three financial mechanisms were established to support distressed countries: the European Financial Stabilisation Mechanism, the European Financial Stability Facility, and the European Stability Mechanism. These ad hoc LOLRs averted the default of Ireland, Portugal, and possibly Spain, and also enabled Greece and Cyprus to remain solvent until (respectively) February 2012 and June 2013. Thus these different types of financial assistance agreements have reduced the number of sovereign defaults, postponed the debt restructuring of less creditworthy countries, and thereby enhanced the perceived accuracy of sovereign ratings.
In short, we observe that the accuracy of sovereign ratings depends on (i) the governments’ respective abilities to line up LOLR-sponsored bailouts and (ii) the CRAs’ alacrity at downgrading a country that will inevitably default. The second aspect is crucial: it highlights the difficult trade-off between the accuracy and the stability of sovereign ratings.
- Accuracy and Stability
Sovereign ratings, which are revised at least once every year, indicate the probability that a country will default in the medium term. The implication is that ratings should be far more stable than most market-based indicators, which fluctuate daily and give excessive weight to the capacity of a government to obtain short-term liquidity. This particular approach to rating was systematized in Moody’s ‘through the cycle’ rating policy.
Not long ago, through-the-cycle rating methodologies were advocated by both investors and regulators. Stable ratings were said to be more convenient for the investors whose investment eligibility guidelines were based on ratings cutoffs. Stability was also promoted by regulatory bodies as a countercyclical tool. Yet, the Eurozone debt crisis revealed that through-the-cycle rating strategies are poor predictors of default. Rather, they seem better suited to low-intensity financial crises – and even then, since ex ante assessments are both difficult and unreliable, accurate rating often relies on LOLR intervention. Cumulative default rates and accuracy ratios would have been less favorable (i.e., more realistic) if CRAs had not sharply reduced the ratings of any country whose default likelihood was surging.
These considerations suggest that sovereign ratings are too high when a major crisis hits the country or are downgraded too slowly to accurately reflect the distressed borrower’s current credit position. Either account would explain why credit ratings persistently lag behind market indicators in times of high risk aversion. Concretely, the inertia of sovereign ratings during the first months of the Eurozone debt crisis raises doubts about the impartiality of CRAs.
- Impartiality and Credibility of Sovereign Ratings
The widespread use of credit ratings in regulatory standards and the business model of the credit rating industry have inflated credit ratings and undermined their impartiality. Sovereign ratings, too, have been affected by this noxious tendency – albeit to a much lesser extent than structured finance (SF) ratings.
Three cases can be used to illustrate the compromised and inflated sovereign ratings issued by the top three CRAs. First, consider the sample of countries that were either upgraded to investment grade or downgraded to speculative grade during the period 1 January 2001 to 1 January 2015. Fitch was most frequently the first firm to upgrade but least frequently the first to downgrade. For instance, Fitch downgraded Greece’s rating to the speculative-grade category 7 and 8.5 months after Moody’s and S&P, respectively. This lenient rating policy may have been implemented by Fitch to appease European policymakers and to preserve the agency’s market share. A second illustration is the puzzling discrepancy observed between S&P’s opinions and ratings of Italy, Portugal, and Spain in 2009: the high ratings assigned (resp. A+, A+, and AA+) were in stark contrast to the pessimistic analyses contained in S&P’s reports. One can reasonably assume that S&P was already well aware of the nascent debt crisis – and its possible magnitude –, but feared early downgrades might result in a self-fulfilling prophecy. Moody’s rating policy regarding Greece in April-May 2010 also reflected the inconsistency of sovereign ratings. Moody’s refrained from downgrading Greece to speculative grade because Greek government bonds would then have become ineligible for Eurosystem credit operations. Moreover, the agency announced that it would not revise Greece’s rating until details of the Eurozone/IMF program were released. This episode, which evidenced the countercyclical action of CRAs, was overlooked when European legislators began working on new regulations to supervise credit ratings.
- The 2013 EU Regulation on CRAs
In addition to deleting references to credit ratings in existing regulatory guidelines and to establishing a civil liability regime for CRAs – two major and beneficial reforms – Regulation (EU) No.462/2013 of the European Parliament and of the Council of 21 May 2013 amending Regulation (EC) No.1060/2009 on credit rating agencies contains several provisions related to sovereign ratings.
First, sovereign ratings must be reviewed at least every six months. This requirement is largely moot because CRAs usually initiate a rating review whenever new information is considered likely to affect the issuer’s creditworthiness. Credit rating agencies are expected to publish a calendar at the end of December for the following 12 months, thereby setting the dates for publication of sovereign ratings. This provision could have the pernicious effect of concentrating speculation in the run-up to the announcement date. Second, CRA reports are required to explain ‘all the assumptions, parameters, limits and uncertainties and any other information taken into account’ when determining sovereign ratings; at the same time, those reports cannot contain policy recommendations, prescriptions, or guidelines. This latter requirement was an overreaction to the sharp yet slow-in-coming downgrades that occurred in 2010–2011, which were perceived by Eurozone governments as a crime de lèse-majesté. This incident is precisely why the Commission and the Parliament supported the creation of a ‘European credit rating agency dedicated to assessing the creditworthiness of Member States’ sovereign debt.’
A weakness of Regulation (EU) No.462/2013 is its view of sovereign ratings as fundamentally procyclical indicators. That misguided diagnosis prevents the legislation from adequately addressing the main issues: ratings inflation and their consequent lack of credibility.
- Restoring Credibility to Sovereign Ratings
Credibility is based on impartiality, accuracy, and consistency. Now that references to credit ratings in US and European regulatory guidelines have been removed, CRAs’ impartiality should be restored. As a result, the following recommendations aim to eliminate inflated ratings and to improve accuracy metrics by enhancing the methodologies currently employed for sovereign ratings.
First, the defaults of Argentina and Greece in (respectively) 2001 and 2012 demonstrated that exchange rate sustainability is a key determinant of sovereign creditworthiness. An overvalued currency undermines competitiveness and renders a country increasingly dependent on imports. As a result, any shift in the exchange rate regime or policy must be scrutinized. Second, it is crucial to assess the sustainability of GDP growth, because the higher the leverage in a boom period, the more severe the recession that follows. A substantial increase in the ratio of national wealth to national income – or stock market indices and housing prices growing two or three times as fast as GDP – for several consecutive years may be evidence of widespread speculation on certain types of assets. Third, CRAs should not rank as investment grade the debt of countries whose World Bank governance indicators are well below world or regional averages.
Another set of factors that merits incorporation into sovereign rating methodologies is closely related to the ins and outs of financial globalization. The increasing size and riskiness of the financial sector in OECD countries have worsened moral hazard since the 1980s. Although CRAs examine contingent liabilities, agencies provide little information about the risk that a central government might be constrained to bail out a major private entity. In fact, the notion of moral hazard is absent from the three CRAs’ methodologies, and only S&P mentions “bailouts”. Yet, in the current era of too-big-to-fail banks, a sovereign borrower’s credit position does depend in part on major financial institutions’ risk-taking strategies. Until bail-in mechanisms are tested and commonly employed, unexpected bailouts constitute more than a potential burden for a government’s public finances and therefore dictate that sovereign ratings be lowered.
A final recommendation is for CRAs to strengthen the consistency of their through-the-cycle methodologies. It is puzzling to observe that Fitch, Moody’s, and S&P are extremely reluctant to lower the ratings of countries whose fiscal and financial position is durably below the level implied by their sustained GDP growth. During 2003–2007, Greece was unable to reach a primary fiscal surplus despite an average annual GDP growth rate that exceeded 4%. In 2004, after the newly elected Prime Minister Kostas Karamanlis revised the fiscal deficit for 2003 from 1.7% to 4.6% of GDP, both Fitch and S&P downgraded Greece by one notch, but Moody’s made no change. It is clear now that these responses were too lax. The Greek government’s announcement was more than a short-term fiscal slippage: it put into question Greece’s fiscal consolidation and transparency in the medium term. This episode should serve as a good lesson for CRAs seeking to distinguish between cyclical and structural shocks.
 For example, it is disturbing to note that Libya was rated A– and BBB+ by S&P and Fitch (respectively) until February 2011.