Credit rating agencies have come under increased scrutiny since the financial crisis. Their failure to recognise the threats to the financial system prior to the crisis coupled with their steady downgrading of European sovereign debt has led to much criticism, especially from European politicians and economists. This Forum examines the major agencies’ influence, independence and performance and explores whether a publicly funded European agency would improve the situation.
Did Rating Agencies Boost the Financial Crisis?
Most observers, journalists from the yellow-press to trade journals, politicians and even economists feel absolutely confident that the rating agencies (henceforth RAs) bear a formidable responsibility for boosting the financial problems of several peripheral European countries into liquidity and even solvency crises. The three big RAs are regarded as all-powerful, mysterious, ignorant, corrupt and unregulated. A specialised European rating agency is demanded, or at least some form of regulation and control of the incumbent agencies. The professional literature, however, is more differentiated, at least as to the rating of sovereign risks. The recent financial crisis with the downgrading of Greece and, to a lesser extent, of Ireland, Portugal and Spain, affords an opportunity for a further test of the validity of these public charges.
This paper starts with a review of the existing literature. It will then provide some information about the rating market and on the pattern of sovereign rating. This is followed by a case study of the RAs’ justifications for downgrading Greece; Greece has been selected as the problems culminated in this country, and Moody’s provides most of the material, as it is the only agency providing a full set of press releases with explanatory statements. The paper than investigates whether the results found for Greece are a special case or if they are typical for the other three countries as well. Finally, political aspects are discussed and conclusions drawn.
Contrary to public opinion and more or less in accordance with the previous literature, the paper finds that the agencies’ sovereign ratings – the paper does not deal with ratings of securities or banks – follow the market rather than lead it, that their bias tends towards the optimistic side, and that they are clearly exposed to the characteristic forecasting errors: pro-cyclicality, turning-point mistakes, underestimation of changes and incapacity to deal with surprises (shocks). If they do have power, it is power delegated to them by policy and regulation.
Rating Agencies’ Image and Existing Evidence
The three big RAs, Standard and Poor’s, Moody’s and Fitch have a rather bad image. Their ratings are considered to be irresponsible, the main cause of the financial crisis in Greece and, to a lesser extent, of the financial problems in Ireland, Portugal and Spain. By downgrading they are said to have unjustifiably blown up the countries’ existing problems, driven up market rates, prohibited the countries’ access to financial markets and undermined the rescue operations of the IMF and the EU. A US Congress Inquiry Commission emphasised in spring that RAs (and investment banks) had played a major role in causing the world financial crisis. O. Rehn, EU Commissioner responsible for economic and monetary affairs, demanded a basic reform of the rating sector in March 2011, and C. Juncker, head of the Eurogroup, expressed his “great surprise” as to the timing of the downgrading of Greece and Spain at about the same time. R. Brüderle, floor leader of the FDP in the German Bundestag, bemoaned that RAs by “self-confident raising or downing of the thumb can send whole countries into the abyss”. The German bank analyst F. Hellmeyer criticised the agencies’ “irresponsibility”: “They exacerbate the mood against the reform countries. They undermine the incontrovertible success of the reform steps. They ignore the success of the reform steps already achieved.”1 The Austrian economist S. Schulmeister argued more sarcastically in June 2011: “Having rescued Portugal, Europe’s elites heralded an end to the problem, Spain and Portugal were not threatened. As if in defiance the RAs announced the potential downgrading of Spain and Italy a few days later.2 The Austrian newspaper “Die Presse” focused the publicly observed state of affairs in the headline: “Trembling with fear of the rating agencies.”3
The RAs’ dubious image may have several sources. First, there is the old tradition of mankind to punish the bringer of bad news: the agencies’ image was less controversial in periods of upswing; they themselves, however, contributed to this negative image by occasionally threatening to downgrade a country in an attempt to manipulate policy.4 Second is the agencies’ opacity: little is known about the criteria of ratings and with the exception of Moody’s it is even difficult to obtain access to their press releases. Thirdly, a general distrust in firms with thousands of employees and high profits in an oligopolistic market may contribute to the bad image, as may opposition to the condemning of small European countries by big, non-transparent profit-oriented “American” firms. Finally, the ratings agencies’ image may at least partly stem from the pronounced flop of their collateralised debt obligations (CDO) ratings, caused by ignorance (of the complex interactions in this new market) and conflicts of interest.5 Such a conflict, however, is less evident in sovereign rating: contrary to securities rating, for which the debtor, i.e. the issuer, has to pay, in the case of sovereign ratings the potential creditor has to pay for the rating report.6
It should, therefore, not be too surprising that the academic literature suggests a somewhat different picture on the stance of RAs. Reinhart7 emphasises that financial crises are generally difficult to predict, not just by RAs; international interest rate spreads and currency forecasts also perform poorly in predicting such crises. Goldstein et al.8 examine the links between currency and banking crises and changes in sovereign credit ratings by Institutional Investor and Moody’s. Neither rating agency was able to predict banking crises, but Moody’s sovereign ratings have at least some low predictive power for currency crises. Reinhart9 extended the study to 46-62 economies and included S&P ratings. She finds that Institutional Investor’s ratings perform worse than indicators of economic fundamentals in predicting both currency and banking crises; the other two agencies do not perform much better. Downgrades in credit ratings usually follow currency crises. Reinhart concludes that sovereign credit ratings tend to be reactive and procyclical, particularly those for emerging market economies, in which both the probability and the size of a downgrade are significantly greater. Rating agencies could do better in predicting defaults if they incorporated indicators of vulnerability (see below).
These conclusions are consistent with the results of a study of the interwar ratings of foreign debt issued at the New York Stock Exchange by Flandreau et al.10 Similar to the case today, the performance of RAs was not exceptional. Ratings seem to have reacted to similar indicators to those found in modern studies and provided little guarantee against default during a financial crisis. RAs generally did not exhibit forecasting capacities superior to those embedded in available market prices.
The relation of ratings to the yield spreads of sovereign bonds is controversial. While Reinhart11 provides evidence of rating changes following changes in spreads, Cantor and Packer12 find that sovereign ratings have an immediate impact on market pricing for non-investment-grade issues; Reisen and Maltzan13 found that sovereign ratings Granger cause spreads for imminent upgrades and actual downgrades, but that the market response is not sustained for more than one day.
The academic literature on sovereign ratings thus provides some evidence that the ratings tend to follow, rather than precede, market developments and, contrary to public opinion, limited evidence exists that the markets would have reacted differently without the agencies’ sovereign ratings. Neither the establishment of a European rating agency14 nor some form of regulation or control of the agencies could, therefore, bring any advantages: the controllers would be subject to the same forecasting errors. No self-interest in biased ratings or any conflict of interest is apparent in sovereign rating. This, however, is not true of the agencies’ main business, securities rating, which according to American (SEC) as well as European (Basel) rules is relevant for the regulation of the finance industry. In this line of business, agencies effectively face a conflict of interest between acquiring customers and strictness of evaluations: the rating should protect the buyers of the CDOs, but the issuer pays for the rating and the RAs frequently advise him as to the composition of the tranches. Contrary to sovereign rating, where little evidence exists that RAs deliberately manipulate, the evidence is less clear for bond rating. Benmelech and Dlugosz15 found in a set of almost 4000 tranches of Collateralised Loan Obligations (CLOs) that four out of five were AAA-rated, while the average credit rating of the collateral was B+; not surprisingly, the value of these AAA-rated securities fell by 70 per cent in 2007 and 2008.16 The uniformity across CDOs in the sample suggests that most issuers were using the RAs’ model (unknown to the public!) to target the highest possible credit rating at the lowest cost. In a subsequent study Benmelech and Dlugosz17 detected that tranches rated by only one agency were more likely to be downgraded – a finding consistent with issuers “shopping” for the highest ratings available from the RAs. This is consistent with Adelson et al.18 who found that Asset Backed Securities rated by S&P alone were more likely to be downgraded and that tranches rated by both S&P and Moody’s were least likely to default. Much of the criticism of sovereign ratings, therefore, may result from a transfer of the potential bias of security rating. This, however, does not mean that the agencies’ assessments are correct: the survey of the literature has provided examples for both sovereign and securities rating, and Elsas and Mielert19 recognised that Standard & Poor’s ratings of European companies were unable to reflect important and fundamental shocks, or only with quite a lag. The RAs’ ability to react to new information does not appear to be outstanding. It should not be forgotten, however, that this is typical for forecasts, and forecasts generally tend to be erroneous.
The Structure of the Rating Market
The rating market comprises ten firms currently registered as NRSROs (Nationally Recognised Statistical Rating Organisations) by the US Securities and Exchange Commission (SEC).20 It is dominated by the Big Three: Standard & Poor’s, Moody’s and Fitch, with a combined market share above 90 per cent.
Standard & Poor’s history dates back to 1860. It was acquired by The McGraw-Hill Companies, Inc. in 1966 and comprises S&P Indices, S&P Equity Research, S&P Valuation and Risk Strategies, and Capital IQ. In 2010 Standard & Poor’s Ratings and McGraw-Hill Financial had offices in 23 countries and revenues of $2.9 billion. The conglomerate provides credit ratings, indices, investment research, risk evaluations and solutions. In 2009 S&P published more than 870,000 new and revised credit ratings. Its equity research provides investment information on approximately 2,000 stocks.
Moody’s Corporation is the parent company of Moody’s Investors Service, which provides credit ratings and research covering debt instruments and securities, and Moody’s Analytics, which offers leading-edge software, advisory services and research for credit analysis, economic research and financial-risk management. The corporation, which reported revenues of $2 billion in 2010, employs approximately 4,500 people worldwide and maintains a presence in 26 countries. Moody’s Investors Service provides credit ratings, research and risk analysis. The firm’s ratings and analysis track debt covering more than 110 countries, 12,000 corporate issuers, 25,000 public finance issuers and 106,000 structured finance obligations. Moody’s Analytics aims to help capital markets and credit risk management professionals worldwide to respond to an evolving marketplace.
Fitch Ratings is dual-headquartered in New York and London. Founded in 1913 by John Knowles Fitch, it began as a publisher of financial statistics whose customers included the New York Stock Exchange. In 1997, Fitch Ratings merged with IBCA Limited, headquartered in London, significantly increasing Fitch Ratings’ worldwide presence and coverage in banking, financial institutions and sovereigns. Through this merger Fitch Ratings became owned by Fimalac S.A., headquartered in Paris. It represented the first step in Fitch Ratings’ plan to become an alternative global, full-service rating agency. The next step was the acquisition of Duff & Phelps Credit Rating Co., headquartered in Chicago, in April 2000 and of Thomson BankWatch’s rating business in 2001. Today, Fitch Ratings employs over 2,100 professionals in 50 offices worldwide.
Sovereign rating is just one task of the RAs and it is definitely not their cash cow. Rating of securities appears to be much more important, and this is the line of business which increased dramatically: the dollar value of originations of subprime mortgages alone rose from $65 billion in 1995 to approximately $600 billion, and Moody’s profits tripled between 2002 and 2006.21
The RAs’ strong position results exclusively from regulation. Bank capital requirements are tied to credit ratings in the USA in diverse SEC regulations as to portfolio restrictions, access to the integrated disclosure system, trading rules and stopgap measures for newer or unregulated products, and in Europe by the Basel Regulation.22 Even the European Central Bank submits to the RAs’ assessments in specifying the necessary quality of collateral for refinancing banks.23
Time Path of Downgradings
Due to its oligopolistic structure the rating market does not appear overly competitive. Market entry is restricted by the high minimum size and the high front-end cost. The RAs have, in principle, list-price schedules for the issuer, but regular customers may renegotiate fees24, which is highly likely due to the market’s oligopolistic structure.25 Competition may, however, not be restricted to fees but to the strictness of ratings as well. In Benmelech and Dlugosz’s26 sample of ABS CDOs 7 per cent were rated by one RA only, 64 per cent by two and 29 per cent by three. Becker and Milbourn27 found that the increased competition following the material entry of Fitch as a third global player coincided with incumbents’ lower quality ratings of bonds and the ability of ratings to predict default deteriorated. The sovereign ratings of the three incumbents, nevertheless, are highly similar – in cases of both faulty and correct changes in assessment (see Figure 1). But again: the market for sovereign ratings may be different, as it is not the subject of the rating – the countries – which pay for it, but the creditor who needs the full rating report. As a consequence it should be expected that the RAs’ business task and the form of competition support a bias towards an over-critical assessment rather than an over-friendly one. There should be a competition in animadversion rather than in laxity; we shall demonstrate below that this is not the case.
Sovereign ratings use a specific code for assessment which differs between Moody’s and the other two as to the labels used but not as to the content, so that they can easily be compared; for simplicity a numeric equivalent is used in this as in most other studies (Table 1).
|Aaa||AAA||AAA||20||highest quality, smallest risk|
|Aa1||AA+||AA+||19||high quality, very low credit|
|A1||A+||A+||16||upper-medium grade, low credit risk|
|Baa1||BBB+||BBB+||13||moderate credit risk|
|Ba1||BB+||BB+||10||questionable credit quality|
|B1||B+||B+||7||high credit risk, generally poor credit quality|
|Caa1||CCC+||CCC||4||very high credit risk, extremely poor credit quality|
|Ca||CC||CCC||1||highly speculative, potential recovery value low|
A main criticism of the RAs’ assessments is their lack of transparency. They disclose neither the criteria on which their ratings are based nor the methods applied.28 Cantor and Packer’s investigation29 suggests that Moody’s and Standard and Poor’s rating assignments are rather similar and can be explained to a large extent by a small number of well-defined criteria which both agencies appear to weight similarly: per capita income, inflation, external debt, economic development and default history. GDP growth, fiscal balance and external balance lack a clear relation to ratings.30 Moody’s appears to place more weight on external debt and less weight on default history as negative factors than does Standard and Poor’s. Moreover, Moody’s places less weight on per capita income as a positive factor. Reinhart31 suggests that RAs would do well to incorporate indicators of vulnerability that have received high marks from the literature on the antecedents of currency crises. While RAs, she finds, have given much weight to debt-to-export ratios, which have proved to be poor predictors of financial stress, they have given little to indicators of liquidity, currency misalignments and asset price behaviour. Many of these indicators have been shown to be useful in predicting not only currency crises but also debt crises.32
Surprisingly, neither RAs nor the academic studies known to the author deal explicitly with the specific default risk of members of a currency union which lack the instruments of monetary and exchange-rate policy; competitiveness and real wage flexibility stand out as prominent examples. Before the formation of the euro currency area, criteria relevant for participation were broadly discussed under the heading “optimal currency area”. Moreover, attention is paid neither to credit expansion, one of best indicators of imminent financial crises33 nor to the old Domar argument of the interest/growth rate spread.
The Pattern of Sovereign Ratings
As Cantor and Packer34 found for a big sample of sovereign ratings, the incumbent RAs’ sovereign ratings tend to move in lockstep. The correlation is highest for S&P and Fitch for the downgrading of Greece and Spain (0.98) and between S&P and Moody’s for Ireland and Portugal. The correlation between Moody’s and Fitch is somewhat lower for Ireland and Portugal (0.88). In the downgrade period beginning in 2009 Moody’s provided the most optimistic ratings for all four countries, while in the upgrade period of the late 1990s and early 2000s Fitch tended to give the highest ratings for Ireland and Portugal, Moody’s for Greece and S&P for Spain.
The analysis of the pattern of sovereign ratings must distinguish among three periods: the period of upgrading the ratings in the late 1990s and early 2000s, the long period of unchanged ratings up to about 2009, and the period of downgrading, moderately at first but rather aggressively since about mid-2010 (see Figure 1). In the first period all RAs upgraded all four countries until three of them reached the top rating of 20 (AAA/Aaa): Ireland in 1998, Portugal in 2003 and Spain in 2004; Greece reached 16 (A+/A1) in autumn 2002. The upgrade obviously resulted from the accession to the eurozone – 1999 Ireland, Portugal and Spain, 2001 Greece – which eliminated the exchange-rate risk. Default risk, however, remained; the EU treaty explicitly precluded a bail-out clause. The high level of the grade, if not the upgrading, must be considered a relatively friendly assessment for Greece and Portugal: both countries did not come near to the criteria carved out by the intensive debate of the 1980s and 1990s as theoretical and empirical requirements for members of an optimal currency area. The OECD35 urged higher real-wage flexibility and wage restraint in general government for Greece and criticised the high and increasing private indebtedness of Portugal, resulting from the low real interest rates.36 The OECD also remarked that both countries had serious structural problems and twin deficits. Despite high growth Greece could not avoid a budget deficit of 4.8 per cent of GDP in 2002 (see Table 2), government indebtedness of 102 per cent of GDP plus some 80 per cent for private households, low household saving, a deficit in the current account of 6.5 per cent of GDP37 and a construction bubble which led to the doubling of property prices since 1998. Portugal had fewer problems with government finance but its current account deficit of 7.7 per cent of GDP revealed similar competitiveness problems. Combined with weak household saving and considerable private indebtedness, this indicated an unsustainable expansion. The other two countries had fewer problems at that time. For Spain the OECD38 complained about the meagre productivity gains and called for housing and labour market reforms and a simpler system of collective bargaining. Ireland suffered from rising housing prices caused by high demand and speculation and by insufficient infrastructure in general;39 in its 2003 report the OECD proclaimed that the era of the Celtic Tiger was over, due to the burst of the ITC bubble and a deterioration of Irish cost competitiveness: a policy change to ensure that both income expectations and public finance adjusted to a slower growth environment would be indispensable.
Indicators Potentially Relevant for Rating
|2002 - max||max - 2010||2002 - max||max -2010||2002 - max||max - 2010||2002 - max||max - 2010|
|Property prices||+35% 2008||-8%||+75% 2006||-36%||+100% 2008||-8%||+65% 2009||0%|
Notes: 1 2007; 2 Share of households with financial problems in 2009 (ZEWnews, June 2011, p. 8); 3 Yearly difference to Eurozone 10 year gov bonds in pps (Eurozone level 1.5%, 4.2%, 3.8%); 4 2008-2009.
Given the problems facing the four countries at the beginning of their eurozone membership, the granting of top grades to Ireland, Spain and Portugal and an upper medium grade to Greece must be considered as a rather friendly assessment and not as proof that the RAs “unjustifiably blew up the countries’ existing problems”, of which politicians and the media have accused them.
The same friendly attitude towards the countries is noticeable for the long period these ratings were maintained – until November 2008 for Greece, until March 2009 for Ireland and until March 2010 for Portugal and Spain – despite the marked deterioration of the countries’ position. Surely the depth and the geographical range of the financial crisis which began with Bear Stearns in March 2008 and accelerated with Lehman Brothers in September 2008 had been underestimated by both economists40 and monetary authorities.41 But the competitive position of all four countries had been deteriorating since at least 2006 (Table 2) and the risks had increased rapidly: Greece’s budget deficit had grown from 4.8 per cent of GDP in 2002 to 7.7 per cent in 2008; unit labour costs had increased considerably faster than in the eurozone and the current account deficit had more than doubled to an unsustainable 14½ per cent. The interest difference to the eurozone average had risen slightly by 0.3 pps. Portugal had managed to keep its budget deficit at about 2¾ per cent of GDP but its competitiveness decreased alarmingly with fast rising unit labour costs and a current account deficit of 11½ per cent of GDP. Spain likewise had increasing public deficits (from ½ per cent of GDP in 2002 to a Maastricht incompatible 4 per cent in 2008), and the increase in the current account deficit from 3.3 per cent of GDP to 9.7 per cent indicated a serious loss of competitiveness. A construction bubble forced property prices up by 100 per cent. In Ireland the current account deficits as well as the budget deficits had started to increase considerably in 2007, property prices soared and the interest spread rose slightly. The ratings were nevertheless modified for none of the four countries for at least three years.
The deterioration not only shows up with hindsight. Already in 2005 the OECD42 warned that Greece’s growth has in part been achieved “at the cost of a sharply widening fiscal deficit to very high levels and a high and rising public indebtedness”; and in 2007: “The clearest sign of macroeconomic tension is an increase in the current account deficit to about 9½ per cent of revised GDP in 2006. In the absence of currency risk, this mainly serves to highlight concerns about a continuing loss of competitiveness, with consumer price inflation running at about 3¼ per cent at the end of 2006, having remained persistently above the euro area average for many years. Relatively high inflation implies low real interest rates, which fuel domestic demand. However, losses in competitiveness may ultimately undermine growth performance.”43 Similarly: “Portugal’s economic performance has deteriorated markedly since 2000, with the slowdown turning out to be more severe and prolonged than in most other OECD countries. This lack of resilience reveals structural weaknesses”.44 In Spain45 “[t]he very rapid rise in household debt and property market prices … could jeopardise macroeconomic stability”, and one year later46 “[t]he long period of virtually uninterrupted strong growth since the early 1990s has ended. This is likely to bring about lasting and profound economic changes. Housing construction is slowing sharply from an unsustainable level, and private consumption is also adjusting to more restrictive conditions in financial markets at home and abroad. … The exposure of the unincorporated, private domestic savings banks – which hold about half of total banking-sector assets and are, as any other bank, under the supervision of the Bank of Spain – is higher than that of other commercial banks”. Ireland was warned by the OECD47 after the IT boom had gone bust that the era of the Celtic Tiger was over and the country had to adjust to lower growth. The survey issued in 200848 had a chapter on “banking on prudence”, pointing out the risks resulting from substantially rising bank lending, increased private indebtedness (from 100 per cent of income to 200 per cent) and high refinancing on the interbank market.
Given these mounting problems, the delayed reaction of the RAs is surprising. Certainly, other forecasters did not do much better in recognising the depth and duration of the financial crisis, but a more critical assessment by RAs could have been expected for two reasons: firstly because RAs’ principal rating task is to carve out the risk to creditors, i.e. to concentrate on downside risks, while academic forecasters should give equal weight to both sides of the risk, and monetary authorities must be especially careful in (published) assessments not to trigger speculative waves. Secondly, it should be assumed that RAs have specific knowledge acquired from rating securities, especially structured securities,49 and banks.
Moderate downgrading started in November 2008 for Greece, in March 2009 for Ireland, and in March 2010 for Portugal and Spain. The downgrading followed the market rather than leading it. The Greek spreads50 (versus the euro area) had reached their minimum from 2005 to 2007 with 0.2 pp and rose to 0.5 pp in 2008, 1.4 pp in 2009 and 6.7 pp in 2010. S&P started to downgrade Greece moderately in January 2009, Fitch followed suit in October 2009 and Moody’s joined in December 2009.
Marked downgrading by several notches followed rather late, concentrated in April 2010 and in spring 2011. The April 2010 downgrading (Fitch on 9 April, Moody’s on 22 April, S&P on 27 April) followed the ECB’s announcement of new collateral rules on 25 March and a considerable rise in spreads: from about 1.4 pp in 2009 to 2.4 pp in January 2010, 2.7 pp in March and 4.3 pp in April. The massive downgrading of Ireland in summer 2010 (Moody’s on 19 July and 5 October, S&P on 24 August, Fitch on 6 October) were preceded by a rise of spreads from 1.3 pp in 2009 to 2.0 pp in June, 3.2 pp in September and 4.7 pp in November. Similar cases of ratings following the lead of spreads can also be found for Spain and Portugal. The correlation between rating changes and interest spreads is 0.98 for Greece and 0.94 for Portugal and Ireland. If the ratings are lagged the correlation is a trifle higher. The RAs thus followed the market rather than steered it, which is consistent with Cantor and Packer.51 In their sample of 49 countries, interest spread rose by 3.3 pp in the 29 days preceding rating changes by S&P and Moody’s, but the ratings had a small impact on the market as well: two thirds of the rating changes in Cantor and Packer’s sample moved spreads in the right direction in the two days around the announcement, especially announcements concerning low grades and announcements by Moody’s. According to Reisen and Maltzan52 the announcement effect disappeared, however, within the next few days.
The evidence so far collected, consistent with the previous literature, suggests – contrary to public opinion – that the RAs’ assessments are rather benevolent and too optimistic and that the downgradings are delayed and follow the reaction of the market rather than direct it. These results are rather puzzling. The natural experiment of the Greek tragedy from unjustifiably joining the monetary union to near insolvency is deployed below to throw some light on this puzzle and investigate the events triggering off RAs’ downgrading. Following that, the experience of the other three countries will be outlined to check whether Greece is a special case. The assessments surveyed are predominately Moody’s, which is the only RA providing detailed justifications of up- and downgrading in its press releases, in its annual reports and in special reports on specific occasions.
Rating Agencies’ Justifications for Downgrading
A Case Study of Greece
After the upgrading by Moody’s and Fitch in October 2002 and by S&P in June 2003, the big three RAs conformed in rating Greece 16 up to late 2004 when S&P and Fitch reduced its rating to 15.53 Moody’s held the rating of 16 up to September 2009, and justified it in its yearly updates with Greece’s membership in the eurozone and the historically high rate of GDP growth, not without emphasising that Greece’s rating was constrained by the high level of debt, uncertainties regarding long-term growth and the need for continued structural reform (February 2006). In its Annual Report54 for 2007 (March) Moody’s expected Greece’s debt ratios to fall more rapidly, but pointed out the risk to the country’s competitiveness due to rising labour costs, a risk reiterated in the annual update of May 2008. Similarly Fitch affirmed its rating (16) in its annual reports of March 2007 and October 2008. On 17 December 2008 Moody’s issued a Special Comment, entitled “Greece: Political Unrest and Credit Crisis Erode the Government’s Fiscal Space”, reacting to the riots that began in Athens on 6 December 2008. It maintained its rating nevertheless: “The current deadlock between the government and the demonstrators is exposing a vulnerability that underpins the relatively low level of Greece’s A1  rating: namely, the inability to build up a social and political consensus needed for reform” (my emphasis, G.T.) To evaluate the “low” rating of 16 (upper-medium grade, low credit risk), unchanged by the negative news about the riots, it must be positioned in the environment of a whole world shaken by the financial crisis, the OECD cautioning against a serious loss of competitiveness since 2005, Greece’s current account deficit of more than 14% of GDP, its budget deficit of almost 8%, government debt approaching 100% of GDP, and “increasing political unrest [which] highlight[s] the fragility of the country’s institutions”.55 The market was less optimistic: interest spreads rose tenfold from about 0.2 pp until 2007 to 1½ pp in late 2008.
S&P was the first RA to downgrade Greece to 14 (one notch) in February 2009, but Moody’s actually upgraded its 16 rating from stable to positive at about the same time. The increasing political unrest, “current deadlock ... inability to build up a social and political consensus needed for reform” recognised in December’s special comment were apparently not regarded as a surprise. Fitch, similarly unimpressed by the inability to form a political consensus, affirmed its 15 rating in its annual report of May 2009. The political unrest did not influence market spreads negatively either: spreads returned to a minimum of about 1 pp in August. In autumn things began to change slowly. Starting in October, interest spreads began to soar. Fitch downgraded Greece to 14 on 22 October 2009, and Moody’s put its 16 rating under review for a possible downgrade due to “a sharp deterioration of public finances relative to the previously reported estimates … the projected deficit for 2009 has been revised upwards, potentially to as high as 12.5%. This is more than twice the level of the previous forecast and substantially higher than the eurozone average. The magnitude of the revision perpetuates Moody’s longstanding concerns about the transparency and reliability of official statistics in Greece.” Moody’s detected a risk in the Greek economy’s large competitiveness deficit. Nevertheless, in a report on 2 December 2009 entitled “Investor Fears of Liquidity Crisis in Greece are Overdone”, Moody’s emphasised that investors’ fears that the Greek government might be exposed to a liquidity crisis in the short term were misplaced; its concerns were related to the erosion of the economy’s potential over the long term, not to short-term liquidity risks. Immediately afterwards, however, Fitch (8 December) and S&P (16 December) downgraded Greece to 13, and Moody’s downgraded the country to 15 with a negative outlook.
In January 2010 Moody’s considered the Greek government’s Stability and Growth Programme as consistent with its current A2 rating (15), but cautioned that Greece’s longstanding problems meant that implementation was far from assured. The concerns about the ability of Greece (as well as Portugal and Spain) to roll over their existing debt and finance their ongoing budget deficits, however, had so far not been substantiated by hard evidence; Moody’s disclosed a month later in its Special Comment “Portugal & Greece: Contagion or Confusion?” that the market spreads in Greece, Portugal and Spain nevertheless suggested much larger credit risk differentiation than was indicated by their ratings. Moody’s maintained its rating anyway. Six weeks later, on 26 March, the RA welcomed the ECB’s announcement that it would maintain its minimum credit rating threshold for collateral at Baa3 (11), as this reduced the liquidity risk for Greece. By contrast, the process of mobilising external assistance had, according to Moody’s, been more factious than anticipated. The key credit question was whether, over the coming weeks and months, market confidence would be strengthened by the EU and IMF’s support package. Market confidence apparently reacted negatively: on 9 April Fitch downgraded Greece by two notches to 11, owing to intensifying fiscal challenges, increasingly adverse prospects for economic growth and increased interest costs. The spread had risen to about 4 pp at that time. Moody’s downgraded Greece to 14 two weeks later (22 April) and placed it under review for a further possible downgrade. It identified a significant risk that debt might only stabilise at a higher and more costly level, that the factious mobilisation of external assistance had made it significantly more difficult for Greece to maintain its debt metrics within the A range (14-16), and that the revision of Greek debt and deficit statistics on 22 April 2010 had further raised the bar for the government to achieve the goals it had laid out in the Stability and Growth Programme. Five days later S&P downgraded Greece to 10, which is “questionable credit quality” or junk status. Analyst Marko Mrsnik said that the rating reflected the political, economic and financial challenges which hindered the reduction of the debt to a sustainable level. Moody’s announced two days later that it expected to complete its review of Greece’s A3 (14) sovereign bond rating shortly after the details of the euro area/IMF programme were unveiled. It noted optimistically that the very high projected level of public debt in Greece was neither unsustainable nor unbearable, and the service of the debt consumed approximately 14% of government revenues compared with 20% to 30% in the 1990s. Stabilising the debt would require a tightening of the primary balance by approximately 12% of GDP in the coming years, which would require considerable commitment and sacrifice by the Greek government and population. Should the mobilisation of external support continue to be factious and/or should the Greek government and people fail to fully deliver on and acquiesce to ambitious policy adjustments, Moody’s indicated that this would inflict significant damage to Greece’s creditworthiness. Moody’s, and to a somewhat lesser extent the other RAs, had been on the optimistic side for quite a while and apparently were hesitant to significantly downgrade Greece. On 14 June, however, it proved unavoidable to downgrade Greece’s rating by four notches to 10: “The Ba1  rating reflects our analysis of the balance of the strengths and risks associated with the Eurozone/IMF support package. The package effectively eliminates any near-term risk of a liquidity-driven default … Nevertheless, the macroeconomic and implementation risks associated with the programme are substantial and more consistent with a Ba1  rating … There is considerable uncertainty surrounding the timing and impact of these measures on the country’s economic growth, particularly in a less supportive global economic environment.”
The second half of 2010 found the ratings unchanged, but interest spreads continued to rise to 8 pp. The next round of downgrading started by putting Greece on a credit watch with negative implications: by S&P on 12 December and by Moody’s on 16 December. Fitch downgraded to its competitors’ rating of 10 on 14 January 2011. Greece’s ability to reduce its debt to sustainable levels was called into question by the RAs, given the substantial upward revision in debt levels, the substantial revenue shortfall in 2010 and, not least, the conditions of ongoing support that would be available to Greece in the event that its market access remained cut off. From this point on, the RAs optimistic and benevolent stance changed markedly. On 2 March Standard & Poor’s warned that it could further downgrade Greece’s (and Portugal’s) debt in the coming two months, depending on the outcome of a crucial European leaders’ summit later in March, and on 7 March Moody’s downgraded Greece’s government bond ratings by three notches to 7 and assigned a negative outlook to the rating. This time the downgrade and the spreads’ rise coincided with one another. The decisive cause of the rating action was the lack of certainty surrounding the precise nature and conditions of support that would be available to Greece after 2013 and its implications for bondholders. Moody’s acknowledged that the IMF and European authorities had expressed very strong support for Greece, provided that the country followed through with this economic programme. However, public statements by European officials suggested that additional liquidity support after 2013 would be conditional on a solvency evaluation, the result of which was considered uncertain by the agency at this point in time. If Greece were viewed as insolvent at this time, there would be some possibility that private creditors would be expected to bear some losses. This was the first time this specific point, crucial for the RAs, was explicitly expressed.
From now on the downward spiral’s speed increased. On 29 March S&P downgraded credit ratings for Greece and Portugal, dropping Greece to an 8, i.e. deeper into junk status. The agency said the country’s borrowing needs were such that it would likely need more help on top of the €110 billion in loans it was receiving from IMF and EU. On 5 May Moody’s placed Greece’s bond ratings under review for a possible downgrade, and on May 20 Fitch downgraded Greece to 756 due to the “high probability that the IMF-EU programme … will cease to be fully funded beyond 2011 … [and that] it is highly unlikely that Greece will be able to regain market access during the remaining life of the IMF-EU programme.” One day later the agency indicated a solution: the exchange of old Greek bonds for new ones (the “Vienna initiative”) would be assessed as the last step leading to default, and not as default proper; the ECB could therefore continue to accept Greek bonds as collateral for refinancing Greek banks. On 1 June 2011 Moody’s downgraded Greece by three notches to 457 with a negative outlook, pointing towards the increased risk that the country would be unable to handle its debt problems without an eventual restructuring, paying creditors less than the full amount or paying later than originally planned. 13 June saw S&P join in downgrading Greece to 4, reflecting “our view that there is a significantly higher likelihood of one or more defaults.” Two days later Fitch matched the other RAs by lowering its rating by three notches to 4 as well. The downgrade reflected a significantly higher likelihood of one or more defaults. S&P stated that the “financing gap has in part emerged because Greece’s access to market financing in 2012 and possibly beyond, as envisaged by the current official EU/IMF program, is unlikely to materialise. S&P believes that some official creditors will see restructuring of commercial debt as a necessary condition to such additional funding. We believe that private sector burden sharing could take the form of a debt exchange offer or an extension of debt maturities. In our view, any such transactions would likely be on terms less favorable than the debt being refinanced, which we, in turn, would view as a de facto default according to Standard & Poor’s published criteria. In that event, under our criteria, this would result in the rating on the affected instruments being lowered to ‘D’ , while Greece’s credit rating would be lowered to ‘SD’ (selective default).”58 Spreads continued to rise.
Summing up, little evidence can be found that the RAs “exacerbated the mood against the reform countries”.59 On the contrary, they were rather benevolent and hesitated for a long while. They rated Greece between 14 and 16 through autumn 2009 due to its euro membership and high growth, although they noticed the very high level of public and private debt and the advancing loss of competitiveness. They ignored the political unrest that began in autumn 2008 and the deteriorating macro data. The first round of sovereign downgrading by one notch (to 13-15) in autumn 2009 was justified by the longstanding problems, the increasing implementation risk and the fact that the external assistance package was more factious than anticipated. The second, rather late round of downgrading in spring 2010 (to 10-11) was caused by the revenue shortfall and Greece’s increasing lack of ability to reduce debt to a sustainable level, the revision of spurious government statistics and, last but not least, by increasing uncertainties surrounding the precise nature and conditions of external support. The third and strongest round of the downward spiral in spring 2011 reduced the ratings to the junk level of 4. The benevolent and lagging character of the assessments was for the first time replaced by a certain aggressiveness, attempting to explicitly avert certain policy decisions. The RAs threatened to downgrade the country to C (0) to avoid any participation, forced or willingly, of private creditors in a haircut or even an extension of maturities. This threat, however, was only possible due to the ECB’s rules on collateral for refinancing banks. Given these rules and the RAs’ task of protecting creditors by providing information on potential risk, the agencies’ late U-turn is not incomprehensible. Incomprehensible is rather the long period during which the RAs ignored Greece’s rapidly deteriorating position. Apart from insufficient understanding of the mechanisms of a currency union and the no bail-out clause’s effectiveness, traditional forecasting errors loom as the main explanation: pro-cyclicality, turning-point mistakes, underestimation of changes and the potential to adjust, disregard of political elements and an incapacity to deal with surprises (shocks). The abrupt change from underreaction to overreaction was primarily caused by the perception of the late, reluctant and contentious support by the EU-IMF and their unrealistic assumption of a return to market finance within one year. In addition, they had to respond to the negative reaction of the market, identifiable by the steep rise in spreads. Thus the RAs followed the unsteady appraisal of the market, with spreads that were below what fundamentals would have suggested from late 2005 until mid-2006 and considerably above what fundamentals suggested in 2010.60
The reasons given for downgrading deviate somewhat from those Cantor and Packer61 found in their analysis of 49 countries, in which per capita income, inflation, external debt, economic development and default history were seen as the main determinants; GDP growth as well as fiscal and external balances had no influence on the ratings. For the Greek downgrades, one-third of the justifications (9 out of 29) were related to the lengthy and “factious” discussion of the rescue concept and its inadequacy in terms of quantity and length of horizon. The only other reason mentioned at least three times (10 per cent) was the projected level of debt (see Table 3).
Justification for Downgrading
Some Reflections on the Downgrading of Portugal, Ireland and Spain
The conclusions from the case study of Greece are by and large reaffirmed by the assessment of Portugal, Ireland and Spain. Portugal and Ireland were subject to two rounds of downgrades – a small and rather benevolent first one, which disregarded the ballooning problems, and a compensating, somewhat overcorrecting second one which cut them by several notches. RAs again underestimated changes and the potential to adjust and disregarded political problems of adjustment; the assessments proved incapable of dealing with surprises, and their switch to a more critical stance was closely related to the controversial negotiations for external support.
Portugal was rated 18 to 20 during 2001-2005 and held this rating until early 2010. The market, however, was less optimistic: the interest spread relative to the eurozone average rose from about 0 - ¾ pp in 2008 to about 2½ pp in early 2010. The first round of weak downgrades in spring 2010 reflected the failure of successive administrations to consistently limit government budget deficits since Portugal joined the eurozone at its inception. The agencies correctly identified low domestic savings and poor economic competitiveness, related more to low productivity than to high costs per se, as the roots of the country’s low trend growth rate. In spite of budget and current account deficits of about 10% of GDP, the forecasts assumed positive, albeit relatively slow, real economic growth. Downgrading did not exceed two notches at that time. Interest spreads declined after the first round to about 0.25 pps but rose again starting in early 2010 to 1 pp at the beginning of the second round of downgrading in spring 2011. This round was triggered less by market pressure than by political uncertainty, resulting in a government crisis as a consequence of the failed austerity programme. The interim government’s limited capacity to act challenged a timely support by a hesitant EU and IMF. Moody’s optimistically justified the limited April downgrade to 13 (and not lower) by arguing that assistance would be provided by the other members of the eurozone if Portugal needed financing on an expedited basis before it could obtain funds from the European Financial Stability Facility, and that the new government would approach the facility as a matter of urgency. Fitch and S&P (in March and April) downgraded Portugal to 11 in big jumps. S&P argued that the new rules on bailout loans, which take effect in 2013, mean sovereign debt restructuring is a “potential precondition to borrowing” from the future European Stability Mechanism and that senior unsecured government debt will be subordinated to ESM loans. Both aspects, announced after a meeting of European leaders in Brussels on 25 March, are “detrimental to commercial creditors.” This is the same threat that has been described above in the case of Greece. It is interesting to note that in each month from December 2010 to March 2011 at least one RA downgraded: ratings fell from 14-17 to 11-13. The spreads, however, did not rise but instead fluctuated around 3 pp. After closing the EU-IMF support programme in May, market spreads rose to about 6½ pp, and on 5 July Moody’s downgraded Portugal by four notches to 9 with a negative outlook. Moody’s said the motive behind the downgrade was the growing risk that Portugal would require another round of financing from its European neighbours, with funding from the private sector a precondition. This exactly parallels what S&P argued when downgrading Greece to junk.
Ireland had a top rating of 20 from late 1998 to spring 2009. The first round of downgrading in spring 2009 resulted from the agencies’ surprise that the fiscal cost to the government of supporting the Irish banking system would turn out to be significantly higher than they had expected. Unfavourable business cycle forecasts contributed as well. The magnitude of the downgrading by two notches on average should be considered as moderate, given that the OECD had already declared the Celtic Tiger era to be over in 200362, that the current account had deteriorated considerably since 2007, that the budget deficit had increased to 7% of GDP in 2008 and to 14% in 2009, and that the interest spread had gone from negative in 2007 to 0.2 pps in 2008 and about 1.5 pps in spring 2010. Until spring 2010 it shrunk to 1 pp; problems had apparently calmed down. The second round, characterised by multi-notch downgrades63 starting in autumn 2010, followed a tripling of spreads to about 3 pps. It had as its starting-point the crystallisation of bank-related contingent liabilities, the increased uncertainty regarding the country’s economic outlook and the decline in the Irish government’s financial strength – facts, however, which were not new but had existed for quite a while. The only new aspect was the annoying debate, both in the EU and in Ireland, over the €85 billion aid package from European governments and the IMF, ultimately agreed on 28 November and finally accepted by Parliament on 15 December. “The austerity measures could have feedback effects on economic growth, on domestic demand, and that’s something that should be monitored,” Moody’s wrote, but it was “very unlikely” that Ireland would default on its debt. Irish borrowing costs initially rose after the bailout was agreed but declined afterwards. The ratings declined to 11-13 in spring 2011. In April Fitch said that the “Irish economy appears to be nearing stabilisation and the latest efforts to resolve the banking crisis are credible”. The market spreads, however, were still rising and stood at 6½ pps in June 2011.
Spain’s sovereign rating was continually upgraded until all RAs agreed on 20 in late 2003. S&P reduced to 19 in January 2009, but the other two kept the top rating through spring 2010. One-notch downgrades by all three RAs reflected their view that the process of adjustment to a lower level of private sector and external indebtedness would materially reduce the rate of growth of the Spanish economy over the medium term. As in the case of Greece, neither the market64 nor the RAs reflected on the consequences of slower growth for unemployment, which had already climbed to 18% in 2009, and hence for political unrest. Surprisingly, none of the RAs worried about the Spanish housing and construction crisis. The OECD, in contrast, was severely concerned about “construction … slowing sharply towards a level which is unsustainable in the long run and investors and consumers are also adjusting strongly to a marked deterioration in financial conditions”.65 Further downgrading in late 2010 and early 2011 by one notch to 18 and 19 respectively was due to the RAs’ discovery of the long-term impact of restructuring Spanish savings banks. Fitch said that its negative outlook reflected the downside risks to Spain’s sovereign credit profile from a weak economic recovery, banking sector restructuring and fiscal consolidation, especially by regional governments. As in the other countries, the rating reflects the market’s assessment: Spreads stabilised at less than 1½ pps, well below Portugal’s 5½, Ireland’s 6½ and Greece’s 12 pp.
The justifications for downgrading these three countries differ from those given by Cantor and Packer66 as well as from those found for Greece. The fiscal cost of restructuring the financial sector was by far the dominant triggering element for Ireland (11 out of 21 statements) and one of three important elements for Spain, the other two being its budget deficit and low growth (in each case 3 out of 12). For Portugal’s downgrades, the country’s lack of fiscal discipline (5 out of 17), low growth expectations (4) and structural problems were mentioned most frequently.
Policy Implications and Conclusions
The study has dealt with the sovereign ratings of the big three RAs, not with their other, quantitatively more important business of rating securities. While the literature offers some evidence of manipulating the rating of securities, especially structured finance obligations, few indications of excessive sovereign downward rating could be detected. June 2011 was the only exception, when two RAs threatened to downgrade Greece and potentially Portugal to insolvency status if private creditors participated in the losses, whether forced or willingly. Even in this case, the problem is less the level of the rating so much as the attempt to force policy. That the RAs would display different behaviour in the two markets is not implausible. In security rating the debtor, who is naturally interested in the highest possible rating, pays for the rating, and the competition for customers cautions the RAs not to be too strict. A clear conflict of interest exists. By contrast, in sovereign rating the potential creditor pays (for the full-length rating report), and he is interested in a stricter rating. One should therefore expect overcritical sovereign ratings, and this is what media and politicians blame the RAs for.
The outcome of this study suggests that just the opposite is true: in most cases the ratings had been benevolent rather than overcritical. However, nothing suggests the RAs would benefit from such a benevolent stance; more likely they were caught by the problems and uncertainties of forecasting resulting from the (at the time) barely understood mechanisms of a currency union without a political union. Uncertainties as to the real effectiveness of the no bail-out clause appear to have combined with traditional forecasting errors – pro-cyclicality, turning-point mistakes, underestimation of change and the potential to adjust, disregard of political elements, and an incapacity to deal with surprises (shocks). The abrupt change from underreaction to overreaction was caused by the late, reluctant and contentious support from the EU and the IMF and their unrealistic assumption of achieving market financing within one year. Furthermore, RAs appear to have given too much weight to financial indicators and not enough to the underlying real problems.
Greece’s dramatic decline in the financial markets was caused only to a very limited extent, if at all, by its ratings. The main reason was a series of policy failures67, by both the Greek and the EU governments. Greece’s unjustified accession to the monetary union, problems in solving structural deficiencies and political discrepancies gave rise to burgeoning debt levels. When it came out that the government had falsified statistics before and after joining the monetary union and the country’s finances were even worse than perceived, interest spreads rocketed and market access was denied even quantitatively. The EU and IMF rescue operation was seen as a disaster for several reasons. It took extremely long to come to an agreement, and as a result the next rescue operation became unavoidable before the first one had even been completed. The support was a political compromise68, offering too little to survive and too much to die immediately. Contrarily, the austerity program was overambitious and so strict that political unrest, even a fall of the government, was an almost certain consequence.69 Finally, the programme’s horizon was far too short; it was absolutely unrealistic to assume that Greece could attain full market access within one year.
The policy failure received its drama through the institutional failure of having granted regulatory power to (private) rating agencies. One may debate whether it is appropriate that rating agencies, by rating securities, control private banks’ portfolio compositions, given the assessments’ limited accuracy. But it is strange indeed that the European Central Bank, which places great emphasis on its political independence, allows its refinancing conditions for banks to be controlled by private companies – the ECB’s rules prevent it from accepting 0-rated securities as collateral. The solution, therefore, is not to gain some control over the RAs’ sovereign ratings or even to found a European RA (which is a highly unrealistic proposal) but rather to revoke the RAs’ control of the central bank’s actions.
- 1 http://www.eu-info.de/dpa-europaticker/185562.html, our translation.
- 2 S. Schulmeister: Endspiel um den Euro, in: Die Presse, 18 June 2011, p. 4; own translation.
- 3 S. Riecher: Das große Zittern vor den Ratingagenturen, in: Die Presse, 16 June 2011, p. 15.
- 4 The UK in March 2011, the United States in April 2011, Italy in June 2011 and the threat to downgrade Greece to default status to avoid the integration of private creditors in the financial restructuring, even if voluntarily.
- 5 The rating should have been to protect the buyers of the CDOs but the issuers paid for the rating and the rating agencies advised them as to the composition of the tranches.
- 6 This – the avoidance of free-riders – is the main reason why so little information is disclosed by the rating agencies.
- 7 C. Reinhart: Default, currency crises, and sovereign credit ratings, 2002, http://mpra.ub.uni-muenchen.de/13917/.
- 8 M. Goldstein et al.: Assessing financial vulnerability: An early warning system for emerging markets, Institute for International Economics, Washington DC 2000.
- 9 C. Reinhart, op cit.
- 10 M. Flandreau et al.: To err is human: rating agencies and the interwar foreign government debt crisis, BIS Working Paper 335, 2010.
- 11 C. Reinhart, op.cit.
- 12 R. Cantor, F. Packer: Determinants and impact of sovereign credit ratings, in: FRBNY Economic Policy Review, Vol. 2, No. 2, 1996, pp. 37-53.
- 13 H. Reisen, J. v. Maltzan: Boom and bust in sovereign ratings, OECD Development Center Working Paper 148, 1999.
- 14 This argument apparently disregards the fact that Fitch has in fact a European (French) owner; a large amount of the research is done in the United States, however, where the dominant market is located and where most of the agency’s customers reside.
- 15 E. Benmelech, J. Dlugosz: The alchemy of CDO credit ratings, NBER Working Paper 14878, 2009.
- 16 M. Pagano, P. Volpin: Credit rating failures and policy options, CEPR Discussion Paper 7556, 2009.
- 17 E. Benmelech, J. Dlugosz: The Credit Rating Crisis, NBER Working Paper No. w15045, 2009.
- 18 Cited in ibid.
- 19 R. Elsas, S. Mielert: Rating Validation based on Shocks to Firms’ Credit Quality, Ludwig-Maximilians-Universität, Working Paper, Munich 2009.
- 20 SEC: Credit Rating Agencies – NRSROs (http://www.sec.gov/answers/nrsro.htm). Under the Credit Rating Agency Reform Act, an NRSRO may be registered with respect to up to five classes of credit ratings: (1) financial institutions, brokers or dealers; (2) insurance companies; (3) corporate issuers; (4) issuers of asset-backed securities; and (5) issuers of government securities, municipal securities or securities issued by a foreign government.
- 21 P. Bolton et al.: The credit ratings game, NBER Working Paper 14712, 2009, p. 1.
- 22 Secretariat of the Basel Committee on Banking Supervision: The New Basel Capital Accord: an explanatory note, January 2001, The standardised approach for credit risk: “Under the new Accord, the risk weights are to be refined by reference to a rating provided by an external credit assessment institution (such as a rating agency) that meets strict standards.”
- 23 Guideline ECB/2000/7 on Monetary Policy Instruments and Procedures of the Eurosystem (ECB/2010/13), amended 16 September 2010.
- 24 L. White: The credit rating industry: An industrial organisation analysis, in R.M. Levich et al.: Ratings, rating agencies and the global financial system, Boston 2002, Kluwer.
- 25 The SEC found that in a sample of subprime RMBS deals, 12 arrangers represented 80% of the business (P. Bolton, op. cit.).
- 26 E. Benmelech, J. Dlugoszs: The alchemy of CDO credit ratings…, op. cit.
- 27 B. Becker, T. Milbourn: How did increased competition affect credit ratings, NBER Working Paper 16404, 2010.
- 28 M. Pagano, P. Volpin, op. cit.
- 29 R. Cantor, F. Packer, op. cit.
- 30 According to K. Bernoth, B. Erdogan: Sovereign bond yield spreads: A time-verying coefficient approach, European University Vidriana Frankfurt/Oder, Discussion Paper 289, 2010, government finance was a relevant determinant of spreads at the beginning of the Monetary Union and again since 2007 but not in between.
- 31 C. Reinhart, op. cit.
- 32 E. Detragiache, A. Spilimbergo: Crises and Liquidity – Evidence and Interpretation, IMF Working Paper 01/2, 2001.
- 33 C. Borio, P. Lowe: Asset prices, financial and monetary stability: exploring the nexus, BIS Working Paper 114, 2002.
- 34 R. Cantor, F. Packer, op. cit.
- 35 OECD: Economic Surveys: Greece, Paris 2001.
- 36 OECD: Economic Surveys: Portugal, Paris 2004.
- 37 “Over the longer term severe spending pressures are building up and maintaining a sizeable primary surplus will be a challenge.” OECD: Economic Surveys: Greece, Paris 2001.
- 38 OECD: Economic Surveys: Spain, Paris 2001; OECD: Economic Surveys: Spain, Paris 2003.
- 39 OECD: Economic Surveys: Italy, Paris 2001.
- 40 G. Tichy: War die Finanzkrise vorhersehbar?, in: Perspektiven der Wirtschaftspolitik, Vol. 11, 2010, No. 4, pp. 356-82.
- 41 G. Tichy: Why did policy ignore the harbingers of the crisis?, in: Empirica, Vol. 38, 2011, No. 1, pp. 107-30.
- 42 OECD: Economic Surveys: Greece, Paris 2005.
- 43 OECD: Economic Surveys: Greece, Paris 2007.
- 44 OECD: Economic Surveys: Portugal, Paris 2006.
- 45 OECD: Economic Surveys: Spain, Paris 2007.
- 46 OECD: Economic Surveys: Spain, Paris 2008.
- 47 OECD: Economic Surveys: Ireland, Paris 2003.
- 48 OECD: Economic Surveys: Ireland, Paris 2008.
- 49 85% of the dollar value of CDO securities was rated AAA by either Moody’s or S&P.
- 50 Long-term government bonds, Greece versus the countries of the eurozone.
- 51 R. Cantor, F. Packer, op. cit.
- 52 H. Reisen, J. v. Maltzan, op. cit.
- 53 For better comparability the numerical equivalents of ratings are applied (see Table 1).
- 54 The RAs change the ratings when they consider it necessary but write Annual Reports reviewing the respective current ratings; in addition they publish Special Reports when considered appropriate.
- 55 Moody’s: Special Comment, 17.12.08.
- 56 “High credit risk, generally poor credit quality”.
- 57 “Very high credit risk, extremely poor credit quality”.
- 58 S&P, 13 June 2011.
- 59 F. Hellmayer, quoted in: http://www.eu-info.de/..., op. cit.
- 60 H.D. Gibson et al.: The Greek financial crisis: Growing imbalances and sovereign spreads, Bank of Greece Working Paper 124, 2011.
- 61 R. Cantor, F. Packer, op. cit.
- 62 OECD: Economic Surveys: Ireland, Paris 2003.
- 63 E.g. five notches by Moody’s on 5 October 2010.
- 64 The spreads fluctuated around ¼ pp up to the end of 2009, rose to 2 pp by June 2010 and to about 1½ pp in early 2011; since that time they have declined somewhat.
- 65 OECD: Economic Surveys: Spain, Paris 2008.
- 66 R. Cantor, F. Packer, op. cit.
- 67 D. Snower and P. Bofinger quoted in: Economics Newspaper, http://economicsnewspaper.com/policy/german/e-crisis-experts-accuse-policy-failure-35477.html.
- 68 W. Kösters emphasised that the problems had been exacerbated dramatically by the European Council reacting to the markets instead of leading, and P. Bofinger added that nobody should be amazed about the markets being flushed out if the eurozone governments carry out all there conflicts publicly and the ECB threatens doomsday scenarios, in: Economics Newspaper, ibid.
- 69 For example, even IMF officials disputed the possibility of raising €50 billion through privatisation within one year, given the strong recession, the deficits of the objects and the condition of the administration; €15 billion through 2013 would have been more realistic.
Rate the Rating Agencies!
Rating agencies continue to be in the eye of the storm. After being singled out very early on as contributors to the subprime bubble, they have continued to be criticised ever since. In the meantime, however, an EU regulation has been adopted and already amended once, and discussions are ongoing regarding a further amendment. In the USA as well, rating agency regulation was reinforced in the 2010 Dodd-Frank Act. The persistence of the criticism seems to indicate, however, that the problems have not been solved by the proposed solutions.
Rating agencies are considered to be an essential element of a well-functioning capital market. They are reputational intermediaries between issuers and investors, along with brokers, analysts, audit firms, financial journalists and self-regulatory organisations. The rating agency industry originated in the context of the growth of the capital market driven model in the USA. The two major firms, Moody’s and S&P, which jointly have over 80% market share, are of American parentage, although the third largest firm, Fitch, is of European parentage. The “Big Three” occupy 94% of the global market (European Commission 2008); all the other players are small to very small and have only a local presence.
We would argue that EU policymakers should stop criticising the industry and instead apply the regulation in a manner that works (before changing it again). They should urgently work on eliminating references to the CRA ratings in other pieces of regulation and stimulate new market entry (and thus not create a publicly funded agency). Supervisors should not insist on the protectionist elements in the EU regulation but check whether the market is competitive. Rather than allowing the industry to make claims to “free speech”, they should insist on the liability of rating agencies when adopting certain positions.
The 2009 EU CRA Regulation and the 2011 Amendment
The rating agencies were one of the first policy victims of the financial crisis. Even before the collapse of Lehman, a strong consensus had emerged that the industry should be subject to statutory regulation. A consultative document was circulated by the European Commission in July 2008, and a regulation was formally proposed in November 2008. The regulation was adopted in six months, a record in EU policymaking. It institutes a single licence for CRAs in the EU, subject to the supervision of the newly created European Securities Markets Authority (ESMA), with tight conduct of business rules.
The debate on the appropriate policy framework for rating agencies considerably predates the financial crisis, however. Already in the 1997 Southeast Asia crisis, the late reaction of rating agencies to the public finance situations in these countries was strongly criticised. The same applies to the dot-com bubble in 2001 with regard to the ratings of corporations. At the global level, in 2003 the IOSCO, the International Organisation of Securities Commissions, adopted a “Statement of Principles” on the role of credit rating agencies – without much success, apparently. At the EU level, pursuant to an own-initiative report of the European Parliament (Katifioris report), the EU Commission asked the Committee of European Securities Regulators (CESR, the predecessor of ESMA) in 2004 for advice on CRAs. In a Communication published in December 2005, it decided that no legislation was needed for three reasons: 1) three EU directives cover rating agencies indirectly – the market abuse, the capital requirements (CRD) and MiFID directives; 2) the 2003 IOSCO Code; and 3) self-regulation by the sector, following the IOSCO Code. In the meantime, the USA had adopted the Credit Rating Agency Reform Act in 2006, instituting a licence for CRAs.
The 2009 EU regulation1:
- requires CRAs to be registered and subjects them to ongoing supervision;
- defines the business of the issuing of credit ratings;
- sets tight governance (board structure and outsourcing), operational (employee independence and rotation, compensation, prohibition of insider trading, record keeping), and conduct of business (prohibition of conflicts of interest in the exercise of ratings or through the provision of ancillary services to the rated entity) rules for CRAs;
- requires CRAs to disclose potential conflicts of interest and its largest client base;
- requires CRAs to disclose their methodologies, models and rating assumptions. ESMA is mandated to set standards for methodologies and establish a central repository with the historical performance data.
The 2011 amendment gives the unique supervisory responsibility to ESMA, which had been created in the meantime, and imposes similar disclosure requirements upon issuers of structured finance instruments under the US SEC’s Rule 17g-5.2 These include collateralised debt obligations (CDOs) backed by retail mortgage-backed securities, for example, which the industry failed so miserably in rating, having assigned AAA to many of them which later had to be downgraded to junk status.
The policymakers thus have a formidable instrument available to control the industry. They can control the governance and business model. They can check for the competence of the employees and the appropriateness of their compensation. They can examine the existence of conflicts of interest and verify the methodologies applied. However, by the end of August 2011, more than a year after the coming into force of the first regulation, only 10 of the 22 CRAs that applied had been licensed by ESMA, including none of the Big Three so far! ESMA may still be understaffed to deal with its new tasks, but EU policymakers could detach Commission officials to ensure rules can be applied if the sector is so high profile. This would allow them to start making in situ inspections of the Big Three in case of disagreements with a sovereign downgrade, for example. This would also be helpful for new industry entrants.
The Captive Market for CRAs Left Untouched
Several pieces of regulation have created a captive market for CRAs. While the USA has started to eliminate references to CRA ratings in its regulation, the EU has not done so thus far. The draft of the CRD IV, which will implement Basel III into EU law, maintains reference to ratings, and the ECB continues to use ratings to decide upon the haircuts applied on collateral in its liquidity-providing operations. This is not the case in the USA, as it has not implemented Basel II (largely because the Federal Reserve did not want the vast majority of US banks relying on CRAs for setting regulatory risk weights) and the discount window of the Fed is not based upon ratings.
In its standardised approach, to be used by less sophisticated banks, Basel II bases risk weightings for credit risk exposure on rating agencies’ assessments. Basel II was implemented into EU law in 2005 as the capital requirements directive. The approach has not been changed in the Basel III proposals, nor in the EU’s draft implementing Basel III. On the contrary, the reliance on the standardised approach may become even more prevalent, as the internal ratings-based approach has come under much criticism in the post-crisis context.3 Less known is that the UCITS III directive, governing the asset allocation of investment funds, also requires an investment grade rating for investments by money market funds. The sector has already been raising this issue with the authorities for some time, but they have yet to act.
In the USA, the Dodd-Frank Act (June 2010) requires regulators to remove from their rules any references to, or requirements that depend upon, credit ratings. In June 2011, the Federal Reserve Board issued a report to Congress reviewing the subject and identified 46 references to credit ratings, the majority of which appear in capital adequacy measures for banks, such as risk weighting. The Fed concluded by saying it will propose amendments to remove references to credit ratings from its capital requirements and rely on substitute standards of creditworthiness for capital calculations that currently rely on external ratings.4
CRAs thus have a dream market in the EU. They are in the news on a daily basis in the context of the sovereign crisis, and their use is obligatory under EU legislation. It is thus high time that EU policymakers emulate their US counterparts and eliminate the regulatory reliance on private credit ratings.
Open Up the Market
Although it was feared that a formal licence would reduce market entry, this does not seem to be the case so far. Some 22 entities applied for an EU licence, including some unexpected newcomers such as credit insurance companies. ESMA and competition authorities should thus carefully monitor how the market develops and whether the oligopoly of the Big Three diminishes. Two questions emerge in this context: the financing model of the Big Three and the protectionist elements in the regulation. The EU regulation does not address a potentially perverse element of the CRAs’ business model, i.e. the “issuer pays” model. This has been widely criticised as raising enormous conflicts of interests, but no change has been instituted so far, although other models have been proposed.5 The issuer pays model maintains the captive market for the incumbents, as they have the reputation and networks. A clearing house for ratings, whereby an intermediary independently decides who should do the rating, may help newcomers and at the same time reduces conflicts of interest.
Another element in the EU regulation restricts market entry but does so in a global context. The regulation requires credit ratings produced outside the EU to be endorsed by a CRA registered in the EU. It has been argued that this regime will unnecessarily fragment global capital markets. Foreign companies will be less inclined to raise capital in the EU, as they will need a local endorsement of their rating. The regime could also be qualified as anti-competitive, as smaller CRAs without an EU presence may stop rating EU sovereigns and issuers.
A final element to bring more competition to the sector is to increase CRAs’ exposure to civil liability. Increasing the responsibility of CRAs may be especially constraining for the larger firms, as they have a strong market presence and have grossly failed in the past.
- 1 Regulation 1060/2009 of 16 September 2009, OJ 17.11.2009.
- 2 Regulation 513/2011 of 11 May 2011 amending Regulation (EC) No. 1060/2009 on credit rating agencies, OJ 31.05.2011.
- 3 See the FSA’s consultation paper CP11/9 of May 2011: Strengthening Capital Standards 3 – further consultation on CRD3.
- 4 Federal Reserve: Report to the Congress on Credit Ratings, July 2011.
- 5 See K. Lannoo: What Reforms for the Credit Rating industry? A European perspective for an overview. ECMI Policy Brief, October 2010.
Sovereign Rating Actions: Is the Criticism Justified?
Nobody likes to be “downgraded” in any walk of life. It implies a decline or deterioration. When the downgrade relates to the credit standing of a developed country, one might naturally expect a hostile reaction and counter-criticism. In the case of credit downgrades, it is arguably a peculiar scenario whereby a private company has a potentially huge influence on a government’s borrowing costs. Credit rating agencies (CRAs) have faced venomous attacks following their recent downgrades of euro-zone sovereigns, but perhaps the most hostile reception was for Standard and Poor’s (S&P) downgrading of the United States of America from the coveted “AAA” status in August this year.
CRAs have also faced criticism in the recent past with regard to their role in rating structured finance products in the context of the US subprime crisis.1 In that case, they were viewed as guilty of assigning excessively high ratings. In contrast, during the European sovereign debt crisis, the criticism is based on downgrading too quickly and/or too far.2 The debt market should be viewed as segmented; therefore issues relating to structured finance ratings, corporate ratings and sovereign ratings should be considered separately. This article focuses solely on sovereign ratings, with the aim of highlighting evidence on the behaviour of sovereign ratings and analysing whether recent critiques of CRA actions have a sound basis.
The Importance of Sovereign Ratings
National governments are the largest borrowers in capital markets, accounting for more than 60% of debt issued. Sovereign ratings represent a ceiling for the ratings assigned to non-sovereign issuers within a country. This was particularly problematic for several Greek, Irish and Portuguese banks, since their ratings were downgraded to speculative status. In addition, sovereign ratings contribute to the smooth and efficient working of the global sovereign debt market. They also have a direct impact on a sovereign’s cost of borrowing, and this is central to the question of whether CRA actions contributed to a worsening of the European sovereign debt crisis.
The role of CRAs has expanded significantly during the last 20 years, whereby credit ratings are now heavily hardwired into investment processes, financial contracts and regulatory frameworks. Therefore, CRA news releases have potentially systemic consequences.3 The uses of ratings imply that CRAs must manage a tension between the stability of their ratings and their short-term “accuracy”.4 The main implication is that ratings will only be changed when the issuer has experienced a permanent change in creditworthiness. Transient rating actions can be very harmful to the debt issuers and users of ratings, hence CRAs are well aware of the need to avoid reversals of actions.
To mitigate the stability-accuracy tension, credit outlook and watch are supplemental instruments used by CRAs to signal adjustments in their opinion of issuer credit quality. These instruments perform an important economic function, but are far too often overlooked by commentators when considering a CRA’s actions and performance. Many empirical studies have shown that outlook and watch signals are at least as important as actual rating changes in their market impact.5 CRAs influence stock and bond prices based not only on revealing new information but also with a “certification” role, though this is most evident via their use of outlook and watch signals rather than actual rating changes.6
Recent Regulatory Developments
In response to the perceived role of CRAs in the US subprime crisis, several policy actions have already occurred and new legislation has been passed in the United States and Europe. The International Organization of Securities Commissions (IOSCO) revised the Code of Conduct Fundamentals for Credit Rating Agencies in 2008 to address issues of independence, conflict of interest, transparency and competition. Also, a formal regulation on CRAs was approved by the European Parliament and entered into force in December 2009. This requires CRAs operating in Europe to register with the Committee of European Securities Regulators (CESR). The responsibility for the regulation of CRAs was handed to the European Securities and Markets Authority in July 2011. CRAs are now subject to legally binding rules that are based on the IOSCO Code.
Many other G-20 countries have introduced or are in the process of introducing new regulatory oversight for CRAs. Further, the Basel Committee of the Bank for International Settlements reviewed the role of external ratings in the capital adequacy framework, mainly to incorporate the IOSCO Code into the committee’s eligibility criteria, and to require banks to perform their own internal assessments of externally rated securitisation exposure. The Financial Stability Board7 published a set of principles for reducing reliance on CRA ratings in standards, laws and regulations.
CRAs have been accused of precipitating the sovereign debt crisis by downgrading the ratings of eurozone sovereigns too far and too fast. Politicians across the EU have called for further regulation to improve quality and transparency in sovereign ratings. Proposals from European politicians have generated a mixed response, including the notion of a publicly owned rating agency and a suggestion that CRAs should notify sovereigns three days in advance of a rating event (rather than the normal twelve hours). A recent UK House of Lords report8 argues that the criticisms are largely unjustified since rating downgrades reflect the seriousness of the problems faced by euro-zone sovereigns. It also encourages legislative changes to enhance the quality of national statistical data and advocates that sovereigns should cooperate with CRAs to ensure that their ratings are as accurate as possible.
Recent Examples of Sovereign Rating Behaviour
Figure 3 presents the recent rating history for the four countries which have attracted the most attention during the European sovereign debt crisis, namely Greece, Ireland, Portugal and Spain. For this and later analysis in this article, we use a 58-point numerical comprehensive credit rating scale (CCR) incorporating the actual ratings, the outlook and watch status, as follows: Aaa/AAA = 58, Aa1/AA+ = 55, Aa2/AA = 52 … Caa3/CCC- = 4, Ca/CC, C/SD-D = 1, and we add “+2” for positive watch, “+1” for positive outlook, “-1” for negative outlook, “-2” for negative watch, and “0” for stable outlook and no watch/outlook assignments. We use data from S&P, Moody’s Investors Service (Moody’s) and Fitch Ratings for October 2006 to August 2011.
Ratings History for Greece, Ireland, Portugal and Spain
Note: The figure plots the rating levels (including outlook and watchlist) for the four selected countries, using ratings from S&P, Moody’s and Fitch for 1 October 2006 to 31 August 2011.
These plots illustrate differences of both opinion and timing across agencies. It is apparent that S&P has tended to be the first mover in taking negative actions related to these countries until the last few months of the sample. At the time of writing, a significant difference of opinion has prevailed in relation to Ireland and Portugal. Moody’s downgraded these two sovereigns’ ratings to speculative status in July 2011, while other agencies are still rating them as investment grade. Indeed, Japan Credit Rating Agency still rates Portugal at the AA- level. In its press releases for these downgrades for Ireland and Portugal, Moody’s emphasises heightened concerns that they will not be able to borrow at sustainable rates in the capital markets following the end of their current EU/IMF support programmes. Therefore, the current differences of opinion largely rest on evaluations of these countries’ prospects for effective spending cuts, increased tax revenues/compliance, economic growth and support for the banking systems.
It should be noted that despite huge media attention on Spain and Italy due to concerns over possible spillover effects arising from other “peripheral” countries of the eurozone, the rating actions for these countries have actually been modest. For example, S&P’s current AA rating of Spain is only one notch (on the 20-point scale) below its rating of the USA. The same one notch difference between Spain and the USA also applies to Fitch (although one rating notch higher than S&P for both). Similarly, there has been widespread media speculation about the sovereign rating of France due to the exposure of French banks to foreign governments’ debt, yet it has not been subject to a single rating action from any of the largest CRAs. The large and frequent downgrades during the recent time period have been restricted to the countries which have required international financial assistance programmes, i.e. Greece, Ireland and Portugal.
In general, it is clear from Figure 3 that the three largest CRAs often disagree on their ratings of these countries, whereby S&P (Moody’s) mostly assigns the lowest (highest) ratings.9 There is almost no indication of reversals in rating judgements in Figure 3.
Figure 3 also shows that only few negative actions occurred before July 2008, suggesting that the CRAs did not predict the public debt problems of these European countries at a very early stage. For example, Moody’s did not change the positive outlook of Greece (rated at “A1”) until 25 February 2009. However, controversy over national economic data was another element in the unfolding of the crisis. For example, in late 2009, Greece’s budget deficit was revised from a forecast of 6-8% of GDP to 12.7% and further revised to 15.4% in 2010 under IMF/EU supervision. The CRAs were no different to Eurostat, banks and others in seemingly being unaware of the full implications of the off-balance sheet vehicles and other manipulations used by Greece. Some commentators have drawn on the analogy of the Trojan horse in reflecting on how Greece was permitted to join the euro. Subsequently, Greece was effectively “sheltering” within the eurozone and benefitting from far lower bond yields than would have been reasonable if the true picture of its indebtedness were revealed. The effectiveness of Greece’s “commitment strategy” of being within the euro has been drastically undermined since the realisation of its true debt levels.
A large number of academic studies over the last fifteen years have been able to explain the determinants of sovereign ratings and rating changes quite successfully. It should be no surprise that per capita GDP, real GDP growth, government deficits and government debt levels are crucial variables which heavily influence sovereign rating levels and actions. In the most recent literature, these four variables have been identified as having a specific short-run impact on the sovereign rating.10
There are therefore clear links between increasing deficits, increasing debt, and lower economic growth in Europe and the downgrading of sovereign ratings.
The above literature has effectively been seeking to identify the processes within CRAs’ “black box” of rating methodology. Until recently, CRAs have only been obliged to reveal vague information about their methods. They could argue that revelation of details would undermine their business advantage and potentially lead to a significant lowering of barriers to entry to the industry. However, regulatory developments following the US subprime crisis (see previous section) have led to increased transparency and disclosure in many elements of CRAs’ activities. For example, S&P published a revealing report on its sovereign rating methodology and assumptions in June 2011.11
We find no evidence in the academic literature on the determinants of sovereign credit ratings that there is any regional or other systematic bias by the three largest CRAs. One example (see footnote 10) has reported that EU membership tends to lead to higher sovereign ratings from the three largest CRAs, all else being equal. However, S&P’s methodology (see footnote 11) implies that membership of a monetary union could actually lead to certain constraints on the sovereign rating from three perspectives of political score, monetary score and economic score.
One may consider the UK’s retention of its top-ranked sovereign rating as questionable given the country’s budget deficits in the aftermath of the banking crisis. However, the current UK government was well aware of the implications of a possible rating downgrade and set out a drastic austerity package which had avoiding a downgrade as one of its explicit aims. It has so far been successful in convincing financial markets of the credibility of its austerity measures. Its AAA rating was placed on negative outlook by S&P in May 2009 but removed in October 2010. It is argued that the UK is able to benefit from being outside the eurozone in terms of exchange rate flexibility and the expectation that economic growth will consequently be somewhat less affected by the austerity measures.
Competition in Sovereign Ratings
Much has been written about a lack of competition in corporate ratings, where S&P and Moody’s dominate the market. However, the same is far from true in the case of sovereign ratings. Many sovereigns are rated by five or six credible agencies (e.g. with the US Nationally Recognised Statistical Rating Organisation designation). The third largest CRA is Fitch Ratings. One recently quoted criticism of the industry has been that the three major agencies are USA-based. It should be pointed out that Fitch has dual headquarters (New York and London) and that it is majority-owned by Fimalac SA, which is headquartered in Paris. Another clear riposte to the “USA bias” argument was the downgrading of the US rating by S&P in August 2011. The US government rating was also placed on negative watch by Moody’s and by Rating & Investment Information Inc prior to the August agreement on the raising of the US debt ceiling.
There is also a large academic literature on split corporate ratings (i.e. where CRAs differ in their opinions of a corporate’s creditworthiness) and their impact on bond yields. However, it is less widely recognised that split sovereign ratings are actually a frequently observed situation. It is very common for CRAs to disagree on the sovereign rating level and on the timing of rating actions.
In the context of recent calls for setting up a European public-owned rating agency and for general increased competition in the rating industry, we wish to quantify the fact that the current major players in the industry have frequent differences of opinion on sovereign ratings and follow quite different rating policies. To illustrate this, we use a sample of daily long-term foreign currency European sovereign ratings, watch and outlook for the three largest CRAs for the period from September 2000 (when Fitch started using outlook for sovereigns) to July 2010.12 We split the sample into pre-crisis (up to October 2006) and crisis periods.
Table 4 presents summary information on the rating actions. Rows 4 and 7 reveal a complete contrast between the pre-crisis and crisis periods in terms of upgrades and downgrades. Moody’s is notable in its use of upgrades of greater than one notch in the pre-crisis period. In general, it is also often the “first mover” in sovereign rating upgrades (see footnote 12). S&P is the heaviest user of outlook actions in this sample (and more generally) as shown in row 11. This reflects a policy difference across CRAs. S&P sovereign ratings tend to react more to short-term events (through the outlook mechanism) than those of the other two CRAs (see also the total number of signals in rows 15-17). Moody’s ratings tend to be more stable, but can be adjusted by large amounts when the action is taken (e.g. in the cases of Ireland and Portugal mentioned above). Because the ratings outlook tends to be more reactive in the case of S&P, there are more frequent reversals of actions.
Descriptive Statistics of the European Sovereign Credit Data Sample
|(I) Pre-crisis period||(II) Crisis period|
|1||Number of countries||32||33||33||38||40||38|
|2||Number of 1-notch upgrades||16||44||44||4||10||4|
|3||Number of >1-notch upgrades||15||1||3||0||0||2|
|4||Total upgrades (Rows 2 + 3)||31||45||47||4||10||6|
|5||Number of 1-notch downgrades||1||6||6||13||29||24|
|6||Number of >1-notch downgrades||1||0||1||6||4||5|
|7||Total downgrades (Rows 5 + 6)||2||6||7||19||33||29|
|8||Total actual rating changes (Rows 4 + 7)||33||51||54||23||43||35|
|9||Positive Outlook Signals||14||40||30||13||19||12|
|10||Negative Outlook Signals||3||21||16||30||53||42|
|11||Total Outlook actions (Rows 9 + 10)||17||61||46||43||72||54|
|12||Positive Watch Signals||12||0||4||2||2||3|
|13||Negative Watch Signals||2||3||4||11||20||5|
|14||Total Watch actions (Rows 12 + 13)||14||3||8||13||22||8|
|15||Total positive signals (Rows 4 + 9 +12)||57||85||81||19||31||21|
|16||Total negative signals (Rows 7 + 10 +13)||7||30||27||60||106||76|
|17||Total sovereign credit signals
(Rows 8 + 11 + 14 or Rows 15 + 16)
|18||Percentage of signals related to:|
|19||Geographical origin of the credit signals:|
|European non-euro countries||60||99||96||58||111||78|
|20||Prior action preceding rating changes:|
|Outlook signal within 6 months||2||8||10||9||11||13|
|Outlook signal within 7-12 months||3||17||10||1||4||8|
|Outlook signal within 13-18 months||0||5||1||0||3||1|
|Outlook signal within more than 18 months||0||3||2||0||1||2|
|Watch signal within 14 working days||0||1||0||1||9||2|
|Watch signal within 15-90 working days||11||2||8||10||9||3|
|Watch signal within 91-150 working days||1||0||0||0||0||2|
|Total outlook/watch signals preceding rating changes||17||36||31||21||37||31|
|21||% of rating changes preceded by outlook/watch signals (Row 20 / Row 8)||52%||71%||57%||91%||86%||89%|
Note: This Table presents summary statistics of daily long-term foreign-currency sovereign signals for European countries rated by each agency during: (I) pre-crisis period: 21 September 2000 to 30 September 2006; and (II) crisis period: 1 October 2006 to 31 July 2010.
Rows 18 and 19 show a clear shift to rating actions on investment grade and eurozone sovereigns in the crisis period, especially for Moody’s. Rows 20 and 21 demonstrate that CRAs’ outlook and watch signals have provided particularly reliable prior warning of forthcoming rating changes during the crisis period. Around 90% of European sovereign rating actions during the crisis period have been preceded by an outlook or watch signal.
Agreement/Disagreement Across Rating Agencies on European Sovereign Ratings Using 58-point (20-point) Rating Scale
|Agencies||Number of countries||Daily observations||Split % of whole sample||Higher rating from first agency% of split|
|Pre-crisis (September 2000 – September 2006)|
|Moody’s and S&P||30||47152||47.6% (39.1%)||62.5% (66.0%)|
|Moody’s and Fitch||31||48875||45.6% (39.7%)||64.1% (66.4%)|
|S&P and Fitch||32||47743||39.7% (26.3%)||54.7% (54.1%)|
|Crisis (October 2006 – July 2010)|
|Moody’s and S&P||37||33914||56.8% (47.7%)||81.5% (97.4%)|
|Moody’s and Fitch||35||33440||45.4% (36.8%)||79.0% (74.8%)|
|S&P and Fitch||37||35446||44.6% (32.6%)||28.8% (27.7%)|
Table 5 provides more detail on CRA differences of opinion on European sovereign ratings for the pre-crisis and crisis periods. Results are provided based on the 58-point scale (see above) and also for ratings only (20-point scale). When outlook and watchlist are considered (58-point scale), there is clearly far greater difference of opinion (split) between agencies. We believe this is the more accurate picture and hence discuss these results only. Almost 50% of daily observations have a different rating status between each pair of CRAs. Moody’s and S&P have the highest level of disagreement, while Fitch and S&P show the most frequent agreement. Differences between CRAs are accentuated during the crisis period. In this period, there is a strong tendency for Moody’s sovereign ratings to be higher than those for the other two CRAs (see Figure 3). The recent downgrades of Ireland and Portugal to speculative status by Moody’s (see above) are very untypical of the wider sample of European sovereigns. S&P also tends to assign lower ratings than Fitch during the crisis period, which reflects wider evidence of S&P as a “first mover” in sovereign rating downgrades (see footnote 12 for further detail).
Figure 4 presents information on the extent of rating disagreements (based on the 58 point scale). It shows that most disagreements are within +/- 6 points on this scale (or two notches on the 20-point rating scale). However, rating differences can be up to 20 points (five notches on the 20-point rating scale), albeit for a relatively short time. The tendency for Moody’s to assign higher European sovereign ratings than the other two is evident here, as is the tendency for S&P to assign lower ratings than Fitch.
Extent of Disagreements on European Sovereign Ratings (58-point scale) Across the Largest Three CRAs
Note: Vertical axis is the square root of the number of days for which the rating difference persists. Rating agreements are excluded from the plots.
In preparing this article, we have conducted an extensive review of recent credit ratings literature; analysed the behaviour of European sovereign ratings over the last decade; studied recent publications by the ECB, IMF, Bank of England and other institutions; and considered the recent public comments made by economists, politicians and regulators. In seeking to summarise our views, we find ourselves in broad agreement with the main conclusions and recommendations of the UK House of Lords report on sovereign ratings in July 2011 (see footnote 8). We strongly encourage readers to consider this report. Specifically, we agree with the report’s views that:
- market investors must take responsibility for their own decisions and consider a range of indicators;
- EU governments should focus on any structures which give undue weight to CRA opinions;
- the proposal that sovereign ratings should be suspended for countries in international financial assistance programmes is wholly impractical;
- any publicly funded European CRA would lack market credibility;
- proposals to give sovereigns more advance warning of rating actions are badly flawed.
Finally, don’t shoot the messenger!
- 1 E.g. see J. Mathis, J. McAndrews, J.-C. Rochet: Rating the raters: Are reputation concerns powerful enough to discipline rating agencies?, in: Journal of Monetary Economics, Vol. 56, No. 8, 2009, pp. 657-674.
- 2 Some criticisms related to CRAs’ conflicts of interest and the potential for assignment of generous ratings due to the “issuer pays” model are therefore not relevant here.
- 3 Bank of England: Whither the credit ratings industry?, Financial Stability Paper, No. 9, 2011.
- 4 G. Löffler: Avoiding the rating bounce: Why rating agencies are slow to react to new information, in: Journal of Economic Behavior and Organization, Vol. 56, 2005, pp. 365-381.
- 5 E.g. see A. Afonso, D. Furceri, P. Gomes: Sovereign credit ratings and financial markets linkages – Application to European data, European Central Bank, working paper, No. 1347, June 2011; R. Alsakka, O. ap Gwilym: Rating agencies’ signals during the European sovereign debt crisis: Market impact and spillovers, Bangor Business School, working paper, 2011.
- 6 E.g. see IMF: The uses and abuses of sovereign credit ratings. IMF Global financial stability report: Sovereigns, funding, and systemic liquidity, 2011.
- 7 Financial Stability Board: Principles for reducing reliance on CRA ratings, 2010.
- 8 House of Lords: Sovereign credit ratings: Shooting the messenger?, European Union Committee, 21st Report of Session 2010-12, HL Paper 189, London, UK, 2011.
- 9 We provide further specific evidence on the general validity of this point in the next section.
- 10 A. Afonso, P. Gomes, P. Rother: Short- and long-run determinants of sovereign debt credit ratings, in: International Journal of Finance and Economics, Vol. 16, No. 1, 2011, pp. 1-15.
- 11 Standard & Poor’s: Sovereign government rating methodology and assumptions, 30 June 2011.
- 12 Related evidence for all sovereigns can be found in R. Alsakka, O. ap Gwilym: Leads and lags in sovereign credit ratings, in: Journal of Banking and Finance, Vol. 34, No. 11, 2010, pp. 2614-2626.
What If Credit Rating Agencies Were Downgraded?
Ratings, Sovereign Debt and Financial Market Volatility
Over the last three years, since the financial crisis began, the volatility of financial markets has significantly increased. Such increased volatility, if it becomes a structural feature, is to be regarded as a negative phenomenon. Higher volatility is simultaneously both a signal and a catalyst of uncertainty. Growth in uncertainty worsens resource allocation.
From a macroeconomic point of view, the increase in volatility is particularly important when it affects sovereign debt, for at least four reasons. Firstly, government bonds represent a significant share of financial assets. Secondly, they are generally held by small investors, i.e. citizens/voters, so that increased volatility can translate into higher uncertainty in general expectations, with a greater risk of real effects on the economy. Thirdly, volatility in sovereign debt also tends to affect the volatility of securities issued by resident corporations and banks. Fourthly, and consequently, volatility in government bonds can more easily trigger economic policy responses, which further amplify its effects.
Lately, CRAs have actively developed their activities concerning government bonds: as of July 2010, Standard & Poor’s, Moody’s and Fitch were passing their judgments on 125, 110 and 107 sovereign states, respectively.1 In general, the activity of CRAs can be a factor in contributing to the volatility of government bonds. The empirical analysis confirms such correlation: rating news regarding the publication of a rating or a revision of a CRA’s opinion about an outlook is linked to variations in government bond yields and/or spreads for associated CDSs via a number of different aspects.
First of all, negative rating news tends to have a negative effect, while positive news seems to have less relevant consequences.2 Also, the effect of negative rating news has significantly increased since 15 September 2008, the date of Lehman Brothers’ bankruptcy.3 Secondly, the effects of rating news on yields and margins tend to reinforce each other4, the negative consequences of which have also been noted by the European Union.5 Thirdly, a contagion effect among states is at work: rating news tends not only to directly affect the issuance of the object of the communication but the sovereign debt of other countries as well.6 Lastly, there is a persistence effect between one communication and another: the correlation between price and/or margin variations and rating news is stronger if the CRA had already released a statement on the country in question in the preceding month.7
Hence rating news seems to affect the prices and thus yields of government debt securities. But how can the relationship between rating news and volatility be explained? Under which conditions does it have a positive or negative effect on financial markets? What are the implications for regulation?
The aim of this article is to try to provide answers to these three questions by illustrating and discussing the three different explanations that economic analysis can offer regarding the relation between ratings and the volatility of government bonds. The argument is simple: if volatility were only and always linked to new information contained in rating news, the effect on markets would be physiological. But it is possible that volatility may depend on two other sets of reasons: the effect of regulations which incorporate these ratings and the communication policies chosen by CRAs. This would result in Excessive Volatility Risk (EVR), which is damaging to markets and which it would be opportune to eliminate.
The analysis will be mostly based on the literature devoted to CRAs which was developed during and after the 2008-2009 economic and financial crisis. The objective is to provide a better understanding of the relation between rating news and markets after the structural break represented by the crisis. The article is organised as follows: in the next three sections, the three most likely explanations of the correlation between rating news and volatility are assessed. These are the two traditional ones – information view and regulatory capture view – and a new one, the communication view. The concluding section draws the implications of the analysis in terms of a prescription for regulation design.
Rating News and Information Discovery Effect
In general the activity of CRAs, as expressed through rating news, can be a driver of volatility for government bonds. But this per se is not necessarily a problem. Ratings are by their nature procyclical. The role of ratings is to provide, through the publication of an opinion, information to markets on the likelihood that a bond-issuing agent – company, bank or government institution – may renege on its commitments.8
If we are dealing with new information to markets, a rating becomes relevant because it reduces information asymmetry (information discovery)9 so that markets move in the same direction as the opinion expressed (cliff effect)10; in the case of positive rating news, markets reward the issuing government, while the opposite occurs if the judgment is negative. In addition to this, information discovery can affect future behaviour by the sovereign issuer, whose financial and economic policy choices can be either confirmed or modified according to whether the rating is positive or negative (monitoring effect).11
In other words, if a rating offers new information to the markets, it contributes to lower macro credit risk, even if this is done at the cost of increased macro volatility risk. Any rating news conveying new information has a positive externality, since it reduces credit risk, and a negative externality, since it increases volatility risk; but the net effect is positive by definition. The greater the degree to which rating news is relevant in terms of information discovery, the stronger the effect that that opinion will have on markets. But what does the relevance of rating news depend on? Since rating news is an output, it all depends on the inputs that go into its production function.
The activities of CRAs have developed by following the principles of supply and demand. This process, started in 1841 with the first CRA – the Mercantile Agency founded by Lewis Tappan12 – has led to 150 CRAs active throughout the world today13; of these, about 140 are single-country and/or single-sector oriented, while a number varying between five and ten, active in Japan, the USA and Canada, provide rating news on more than one country or industry.14 In the United States, there are ten officially registered CRAs.15 The global market is dominated, however, by the Big Three – Standard & Poor’s, Moody’s and Fitch – with market shares estimated to be 40% each for Standard & Poor’s and Moody’s and 15% for Fitch.16 The number of issuers subjected to ratings by Standard & Poor’s has climbed from 1386 in 1981 to 5860 in 2009.17
The rating news output is born from the fact that on capital markets investors need information about the agents – corporations, banks, public institutions – that issue equities and bonds. There is a demand for information out there which rating agencies are set to meet (information discovery). Ratings are supplied by private firms: they represent an assessment of the probability that the issuer will regularly and completely fulfil its obligations.18 The assessment is subjective and forward-looking; these two characteristics differentiate ratings from accounting reports, which are instead based on historical data and objective criteria.19 CRAs are intermediaries of information20: given information inputs from various sources and technological and human capital at their disposal, they produce information with a higher value added.
If a rating fulfils the function of information discovery, thereby reducing information asymmetry on capital markets, it produces the so-called certification effect on the quality of the security and the issuer.21 The certification effect summarises the net positive externality of rating news: the action of private firms – CRAs – has widespread effects on the efficiency of all markets; it thus produces a public good, since information can be consumed by all without risk of rationing22, and the effect on the prices of bond emissions, volatility included, is a natural consequence of this.
But what does the production of information discovery, and hence the certification effect, depend on? The prime mover is the incentive for CRAs to foster a respected reputation for themselves (reputation-building). When reviewing an emission and/or an issuer, a CRA tries to give the best possible judgment, putting together public and private information on one side and specialised human capital applying the best methodologies on the other. As the reputation of a CRA grows, its rating news is bound to have a larger impact on the market.
There are at least three senses in which rating news offers value added in informational terms. Firstly, CRAs have access to non-public information sources (data inputs).23 Secondly, they have access to higher-quality human capital and technology to handle such data. Thirdly, CRAs have the correct incentives (goal function) to supply a quality product, independently of the point in the business cycle or the nature of the issuer.
However, recent economic analysis has questioned all three of these justifications for the information discovery produced by rating news, especially in the case of sovereign emissions. Doubts originate from the general fact that ratings have proved ineffective on various occasions, starting with the Asian crises of 1997 and 199824; in the case of California’s Orange County default; in the Enron, WorldCom and Global Crossing cases25; and in the defaults of structured finance26, which was identified as playing a significant role in the origin and development of the 2007-2009 financial crisis.27
The ineffectiveness of rating news can have at least three different causes. First, the release of ratings on government debt, particularly if unsolicited, does not enjoy the advantage of coming from privileged information sources.28 Second, it has been questioned whether CRAs do in fact manage to attract the best human capital with respect to other financial firms and institutions, given their salary and incentive structures29, or if such human capital is indeed inadequate30, also concerning methodological choices.
Finally, there can be biases in the conduct of CRAs that lead to systematic distortions in ratings, independently of the issuance and/or issuer involved. Let us list here only a sample of the hypotheses presented in this regard in the economic literature. A first hypothesis is that the economic cycle has an effect on the degree of homogeneity of ratings: CRAs tend to behave similarly during expansionary phases, while they tend to differentiate their opinions during recessionary phases of the cycle.31 Such a finding would counter the argument that the ratings are constructed with cycle-smoothing techniques.32
A second hypothesis is that CRAs modify the level of severity of their assessments in a countercyclical way in order to accommodate issuers who pay for the ratings: in recessionary phases, opinions are more lenient to meet the need of issuers to find a market for their emissions in more difficult macroeconomic conditions.33 In general, the conflict of interest that is intrinsic to the relationship between CRAs and issuers can cause biased incentives.34 Thus the risk of biased ratings can be linked either to the economic cycle or to the business model adopted by CRAs.
Summing up, the information discovery regarding the reliability of sovereign government emissions that rating news should produce is far from assured. In spite of this, we have seen that rating news continues to have important effects on market volatility. Thus, the relevance of rating news may depend on other factors. In this case, the ensuing volatility would be excessive volatility, since we would have to bear the cost of the increase in volatility risk without the benefit of the reduction in credit risk. But what does EVR depend on?
Rating News and Rating-Based Regulation Effect
The EVR of rating news can be explained starting from the fact that ratings are used as an integral part of various types of banking and financial regulation (rating-based regulation).
Ratings – starting with the first initiative in this field by the SEC in 193635 – have been progressively embodied in numerous and significant regulations. There are at least five areas of regulation which have seen the use of ratings: admission requirements to regulated stock markets (again, first implemented by the SEC in 197536); classification of assets in portfolios of institutional and public investors; valuation of assets in securitisation processes; transparency; and in particular prudential oversight, the most pervasive example being the Basel Accords, starting with the 2004 version37 and reaffirmed by the 2010 Basel III Accord.38 Rating-based regulation has developed precisely because of the information discovery role assigned to ratings.39
The embodiment of ratings in regulation has automatic effects on the likelihood of securities and their issuers finding a market, thus becoming a sort of quasi-public licence that affects the success of an emission (licence effect).
There is widespread consensus that the importance of ratings, and thus the relevance of rating news, has greatly increased since rating-based regulation was developed.40 But if the increase in volatility risk, amplified by license effect, were still based on information discovery, the net effect of rating news could still have been considered as a positive externality.
As time went by and doubts grew about the value of information discovery attributable to ratings, the hypothesis has gained ground that the relevance of a rating can itself depend on the role played by regulation no matter what the informational content.41 The licence effect linked to ratings would then be among the causes of the recent financial crisis.42
In other words, the licence effect ends up being independent of the certification effect. In the presence of a licence effect which explains the relevance of rating news, we would have an effect on the volatility of the value of the issuance and/or issuer which was unjustified when the information content of the rating was considered. Theoretically, the more likely the licence effect is, the higher EVR will be; we shall have a case in which, in the presence of inaccurate public information, there are distortions in financial markets.43 In other words, rating news would only cause an increase in volatility risk, without the information benefits that reduce credit risk; we would thus have a net negative externality (a public bad).
Rating News and Communication Effect
The economic literature has yet to explore a third channel that may explain the relation between rating news and volatility: the communication policies of CRAs. It is surprising that this channel has been overlooked until now, in spite of the importance of communication that is intrinsic to the release of opinions by CRAs. In other fields of economics, the analysis of the role of communication in determining the effectiveness of the transmission of information has been significantly developed – think of monetary policy or, more recently, of tasks of macro-supervision assigned to central banks.44 The same type of research needs to be conducted for CRAs. In fact, keeping the level of information discovery constant, it is intuitive and self-evident how the relevance of rating news is linked to communication policy (communication effect), for various reasons.
Firstly, the importance of communication is apparent, starting with the choice of expressing the rating evaluation through letter grades, a synthetic and immediate way of communicating which is comprehensible to all investors no matter their level of financial literacy.45 Secondly, if the economic role of CRAs is that of information intermediaries, the definition of the modalities and timing of their communication to markets is essential if the transmission of the information contained in the evaluations of CRAs is to be effective. Thirdly, the communication policy adopted is even more important in the case of evaluations of sovereign debt issuances, for the reasons illustrated in the introduction. Fourthly, the increasingly important issue of the accountability of CRAs must be considered.46 Discussion of this topic has so far been limited to liabilities linked to opinions expressed by CRAs. But since the effect that evaluations have on markets depends not only on information but also on communication, designing mechanisms of accountability must necessarily impinge on both aspects of the policy adopted by CRAs.
The crucial point is that communication policy is an integral part of information discovery. The greater the extent that rating news contains information discovery, the more the volatility caused by the communication effect will be physiological. Conversely, the more uncertain the content of information discovery, the higher the risk of EVR.
The policy of communication adopted by CRAs can be studied by highlighting at least three different aspects. First of all, the object of communication must be distinguished, which can be either a rating or an outlook. In principle, we can hypothesise that the effect on markets depends upon the type of communication – an evaluation expressed by a rating or a revision of an evaluation as formulated by an outlook. Secondly, the modes of communication must be considered. These can take the form of a press release, a press conference or something else. Thirdly, the timing of communication must be investigated from two points of view: in absolute terms, by distinguishing periodical, institutional communication, which is predictable, from rating news which is not; and in relative terms, with respect to the functioning of financial markets (e.g. whether rating news is communicated when markets are closed or open).
The activity of CRAs has effects on the volatility of yields and margins pertaining to bond emissions by sovereign governments. If the effect on financial markets were to depend exclusively on the information discovery function contained in rating news, the negative effect would be more than offset by the positive effect in terms of more accurate information to evaluate credit risk. However, since the information discovery function of rating news cannot be taken for granted, there is a risk of excessive volatility, linked to the fact that the rating has become integrated within regulation or because of the communication policy adopted by CRAs. From these considerations, two types of conclusions can be derived, which are linked to positive and normative analysis respectively.
From the point of view of the analysis of the association between rating news and volatility, it is important to conduct empirical studies aimed at distinguishing the relative influence of the certification, licence and communication effects. As far as regulatory implications are concerned, the resulting excessive volatility is a negative macroeconomic phenomenon.
If we were to decide that the risk of excessive volatility ought to be eliminated, we would need to act on at least two fronts. On the one hand, rating-based regulation should be disposed of. Over the last few years, there have been numerous calls at the international level to diminish the role of ratings in regulation in the medium term.47 This should be encouraged and accelerated. Delays in this regard would make all the more robust the thesis that the intervention of regulators and politicians on ratings is slow and inadequate due to strong lobbying by the CRAs themselves, especially in the United States.48 On the other hand, when considering proposals for new regulation aimed at increasing the accountability and liabilities of CRAs, the issue of communication policy should be dealt with explicitly and directly.
- 1 International Monetary Fund: The Uses and Abuses of Sovereign Ratings, in: Global Financial Stability Review, October 2010, pp. 85-122.
- 2 H. Reisen, J. Maltzan: Boom and Bust and Sovereign Ratings, in: International Finance, Vol. 2, No. 2, 1999, pp. 273-293; J.C. Hull, M. Predescu, A. White: The Relationship between Credit Default Swap Spreads, Bond Yields, and Credit Rating Announcements, in: Journal of Banking and Finance, Vol. 28, No. 11, 2004, pp. 2789-2811; L. Norden, M. Weber: Informational Efficiency of Credit Default Swaps and Stock Markets: The Impact of Credit Rating Announcements, in: Journal of Banking and Finance, Vol. 28, No. 11, 2004, pp. 2813-2843; R. Kraussl: Do Credit Rating Agencies Add to the Dynamics of Emerging Market Crises?, in: Journal of Financial Stability, Vol. 1, No. 3 2005, pp. 355-385; A. Afonso, D. Furceri, P. Gomes: Sovereign Credit Ratings and Financial Markets Linkages, Working Paper Series, No. 1347, 2011, European Central Bank.
- 3 A. Afonso, ibid.
- 4 H. Reisen et al., op. cit.; A. Afonso et al., op. cit.
- 5 J.M.Q. Barroso: Statement to the European Parliament prior to the Meeting of the Heads of State and Government of the Euro Area, European Parliament Plenary, Brussels 5 May 2010.
- 6 A. Gande, D. Parsley: News Spillovers in the Sovereign Debt Market, in: Journal of Financial Economics, forthcoming 2010; J. Ismailescu, H. Kazemi: The Reaction of Emerging Market Credit Default Swap Spreads to Sovereign Credit Rating Changes, in: Journal of Banking and Finance, Vol. 34, No. 12, 2010, pp. 2861-2873; A. Afonso et al., op. cit.; R. Arezki, B. Candelon, A. Sy: Sovereign Ratings News and Financial Markets Spillovers: Evidence from the European Debt Crisis, Working Paper Series, No. 11-68, 2011, IMF.
- 7 A. Afonso et al., op. cit.
- 8 P. Deb, M. Manning, G. Murphy, A. Penalver, A. Toth: Whither the Credit Ratings Industry?, in: Financial Stability Paper, 2011, No. 9, Bank of England; J. De Haan, F. Amtenbrink: Credit Rating Agencies, in: S. Eijffinger, D. Masciandaro (eds.): Handbook of Central Banking and Financial Regulation after the Financial Crisis, forthcoming 2011, Edward Elgar; U.G. Schroeter: Credit Ratings and Credit Rating Agencies, in: G. Caprio (ed.): Encyclopedia of Financial Globalization, forthcoming 2011, Elsevier.
- 9 M. Pagano, P. Volpin: Credit Ratings Failures and Policy Options, in: Economic Policy, Vol. 25, No. 62, 2010, pp. 403-431; P. Deb et al., op. cit.
- 10 P. Deb et al., ibid.
- 11 A. Boot, T. Milbourn, A. Schmeits: Credit Ratings as Coordination Mechanism, in: Review of Financial Studies, Vol. 19, No. 1, 2006, pp. 81-118; J. De Haan et al., op. cit.
- 12 P. Deb et al., op. cit.
- 13 J. De Haan et al., op. cit.; U.G. Schroeter, op. cit.
- 14 U.G. Schroeter, ibid.
- 15 P. Deb et al., op. cit.
- 16 U.G. Schroeter, op. cit.; the sum of the three reaches 98% in F. Partnoy: Rethinking Regulation of Credit Rating Agencies: An Institutional Investor Perspective, Legal Studies Research Paper Series, No. 09-14, 2009, University of San Diego.
- 17 P. Deb et al., op. cit.
- 18 J. De Haan et al., op. cit.; U.G. Schroeter, op. cit.
- 19 P. Deb et al., op. cit.; J. De Haan et al., op. cit.
- 20 F. Partnoy, op. cit.; U.G. Schroeter, op. cit.
- 21 P. Deb et al., op. cit.; J. De Haan et al., op. cit.
- 22 U.G. Schroeter, op. cit.
- 23 P. Deb et al., op. cit.
- 24 G. Ferri, J. Stiglitz: The Procyclical Role of Rating Agencies: Evidence from the East Asian Crisis, in: Economic Notes, Vol. 29, No. 3, 1999, pp. 335-355.
- 25 F. Partnoy, op. cit.; P. Deb et al., op. cit.
- 26 Committee on the Global Financial System: The Role of Ratings in Structured Finance: Issues and Implications, No. 23, January 2005.
- 27 President Working Group on Financial Markets: Policy Statement on Financial Market Developments, March 2008; Financial Stability Forum: Report of the Financial Stability Forum on Enhancing Market and Institutional Resilience, April 2008; Issing Committee: New Financial Order Recommendations, February 2009; Turner Review: A Regulatory Response to the Global Banking Crisis, March 2009; De Larosière Group: The High Level Group on Financial Supervision Report, 2009.
- 28 B.J. Kormos: Quis Custodiet Ipsos Custodes? Revisiting Rating Agency Regulation, in: International Business Law Journal, Vol. 4, 2008, pp. 569-590.
- 29 C. Hill: Regulating the Rating Agencies, in: Law Quarterly, Vol. 82, 2004, pp. 43-94, Washington University.
- 30 F. Partnoy, op. cit.
- 31 A. Croce, S. Lugo, R. Faff: Rating Alignment, Rating Shopping and Reputation of Credit Rating Agencies: Evidence from the Sub- Prime Crisis, mimeo 2011.
- 32 R. Cantor, C. Mann: Analyzing the Trade Off between Ratings Accuracy and Stability, in: Journal of Fixed Income, Vol. 16, No. 4, 2007, pp. 60-68.
- 33 T. Wang: Can Reputation Concerns Always Discipline Credit Rating Agencies? Evidence from Corporate Bond Issuance Ratings, mimeo 2011.
- 34 P. Deb et al., op. cit.
- 35 U.G. Schroeter, op. cit.
- 36 P. Deb et al., op. cit.
- 37 International Monetary Fund: The Uses and Abuses of Sovereign Ratings, in: Global Financial Stability Review, October 2010, pp. 85-122; P. Deb et al., op. cit.; U.G. Schroeter, op. cit.
- 38 Basel Committee on Banking Supervision: A Global Regulatory Framework for More Resilient Banks and Banking Systems, December 2010.
- 39 U.G. Schroeter, op. cit.
- 40 Ibid.; P. Deb et al., op. cit.
- 41 F. Partnoy: The Siskel and Ebert of Financial Markets: Two Thumbs Down for the Credit Rating Agencies, in: Law Quarterly, Vol. 77, 1999, pp. 619-714, Washington University; id.: Rethinking Regulation…, op. cit.
- 42 F. Partnoy: Overdependence on Credit Ratings was a Primary Cause of the Crisis, Legal Studies Research Paper Series, No. 09-15, 2009, University of San Diego.
- 43 S. Pagratis: Asset Pricing, Asymmetric Information and Rating Announcements: Does Benchmarking on Ratings Matter?, Working Paper Series, No. 265, 2005, Bank of England; F. Allen, S. Morris, H. Shin: Beauty Contests and Iterated Expectations in Asset Markets, in: Review of Financial Studies, Vol. 19, No. 3, 2006, pp. 719-752.
- 44 For a survey see B. Born, M. Ehrmann, M. Fratzscher: How Should Central Banks Deal with a Financial Stability Objective? The Evolving Role of Communication as a Policy Instrument, in: S. Eijffinger, D. Masciandaro (eds.): Handbook of Central Banking and Financial Regulation after the Financial Crisis, forthcoming 2011, Edward Elgar.
- 45 U.G. Schroeter, op. cit.
- 46 P. Deb et al., op. cit.; U.G. Schroeter, op. cit.; F. Partnoy: Rethinking Regulation…, op. cit.
- 47 Basel Committee on Banking Supervision: Enhancements to the Basel II Framework, July 2009; and id.: A Global Regulatory Framework for More Resilient Banks and Banking Systems, December 2010; Financial Stability Board: Principles for Reducing Reliance on CRA Ratings, 27 October 2010; Securities and Exchange Commission: References to Credit Ratings in Certain Investment Company Act Rules and Forms, March 2011; for a survey, see P. Deb et al., op. cit.
- 48 F. Partnoy: Rethinking Regulation…, op. cit.
Misguided Ventures: A Quasi-Public European Union Credit Rating Agency
As speculative attacks against Member States persist, the European Union (EU) is desperately attempting to allay fears concerning the disintegration of Economic and Monetary Union (EMU). Much of the firepower for this onslaught is provided by those very financial intermediaries charged with assessing and communicating the health of euro area economies, namely credit rating agencies (CRAs). Thus, in 2008, the European Commission proposed a series of oversight initiatives which would centralise CRA supervision at the EU level. Endorsed by the High Level Group on Financial Supervision1 chaired by Jacques de Larosière, Regulation (EC) No 1060/2009 (CRA Regulation v1) came into effect on 7 December 2010. Shortly thereafter, the corresponding amendment (EU) No 513/2011 (CRA Regulation v2) was deemed necessary in order to compensate for outstanding issues.
Already, however, there are serious questions about whether these new multilateral measures are sufficient to prevent Europe from being held hostage by the procyclical behaviour of a cabal of private firms: Moody’s Investors Services, Standard & Poor’s (S&P) and Fitch Ratings. In fact, the European Securities and Markets Authority (ESMA) is busy designing the third regulatory framework to supplement these existing CRA regulations. It is due to be announced at the beginning of 2012. But will it be too late to salvage the current configuration of EMU or even be effective in redressing some of the most egregious elements of ratings? Extreme volatility now threatens the eurozone’s largest economies, including France. Not even the United States is immune. In this attempt to correct some of the numerous offences with which EU officials have charged the CRAs, the European Commission published its 5 November 2010 consultation paper on credit rating agencies.2 To reduce an overreliance on external ratings, inject competition into the ratings space and enhance the transparency of the entire process, a structural solution is being entertained in the form of an EU credit rating agency. But is this the right approach?
Grave concern among EU officials is well warranted. Whether it was their haste in downgrading sovereign debt during the 1998 Asian crisis, their inability to foresee the collapse of fraudulent corporate giants like Enron (2001) and Parmalat (2003), or their assignment of high investment grade ratings to dubious subprime mortgage-backed securities, which contributed to the 2008 credit crisis, credit ratings agencies have been implicated in some of the most severe and destabilising financial and fiscal crises of the last two decades. Procyclicality is observed as negative downgrades hinder debt financing, dampen economic growth and thus precipitate further decreases. Irrespective of the ensuing political outrage and promises to correct such abuses, ratings agencies have managed to elude any serious regulation.
Now, as the integrity of the monetary union itself is undermined, it is tempting to believe that the (re)politicisation of this largely depoliticised field of finance will have the desired countercyclical effects and restore stability to beleaguered Member States and financial markets. Yet can a quasi-public EU CRA actually correct some of the imbalances and inconsistencies evident in the ratings space or would it simply exacerbate them? This paper contends that this is a misguided approach that can only infuse more uncertainty about the quality of ratings, heighten the dependence on external forms of assessment and undermine the EU’s authority to manage effectively the sovereign debt crisis. Arguably, the current regulatory framework is inadequate as it fails to address principal problems, such as a fallacious analytics of ratings. But an EU credit rating agency can only entrench such distortions and amplify destabilising “cliff effects” as it cedes further sovereign authority to market forces.
Sovereign rating ranges rest on a judgement – codified and commercialised as the “risk of default” – about “the capacity and willingness” of governments to raise the necessary resources for the timely servicing of their debt obligations.3 Probability of payment must be concomitant with the tolerability of the costs of austerity/adjustment. Yet as the “pain” threshold which a constituency can endure fluctuates according to its changing political economy, it escapes prescient quantification as a probability distribution through the utilitarian calculus of risk. Politics is just too fluid and uncertain to be captured statistically, even in an ordinal range. Sovereigns rarely default in the way that corporates fail. Nevertheless, in their attempt to make the qualitative more quantitative, the “opinions” that Moody’s or S&P issue about this governmental capacity seem to be widely accepted. Coming to terms with this analytics of ratings and properly regulating it is pivotal to an effective EU policy response. Thus far, the EU has been reluctant to regulate the flawed analytics underpinning ratings themselves, which leaves Europe susceptible to further destabilising forces.
Overreliance on External Ratings
In part, the scope and authority of ratings derives from their “certification” role. Institutionalised in regulatory capital requirements, certification is intended to identify whose ratings are appropriate for regulatory purposes in the EU. In order to be eligible as collateral for money market operations, securities typically must have an investment grade for central banks to accept them. Financial contracts and the by-laws of corporations have similar suitability criteria. The European Central Bank’s (ECB) “credit quality threshold” is defined in terms of BBB+/Baa1 in its harmonised rating scale. Of course, as the recent cases of Ireland and Greece demonstrate, minimum credit rating thresholds can be suspended. Now the ECB will accept Greek defaulted bonds as collateral. Given Greece’s tremendous medium-term solvency challenges coupled with Germany’s move to recapitalise its own banks exposed to Greek debt, its default is imminent.
What is very odd, however, is that rather than removing references to or reliance upon ratings, the ESMA registration process merely serves to enhance the status and the legitimacy of rating agencies. Lessons should have been learned from the “Nationally Recognized Statistical Rating Organizations” (NRSRO) designation in the United States, of which there are ten. Only recently has the Rating Agency Reform Act of 2006 introduced criteria detailing what the NRSRO designation actually entails. Prior to 2006, however, certification by the Securities and Exchange Commission (SEC) was quite informal, which actually erected barriers to entry and solidified the duopoly enjoyed by Moody’s and S&P. Ostensibly, this could repeat itself in the European context or dilute the lower tiers of the ratings industry with a slew of relatively “insignificant” rating agencies, the EU CRA amongst them. Such players are no match against goliaths the likes of Moody’s or S&P who, in 2011, rated 112 and 126 sovereigns, respectively. Even Kroll Bond Rating Agency (KBRA) – at 59 sovereign ratings – is not considered as a potential challenger to these global full-spectrum rating agencies. Most new entrants simply appear resolved to carve out niche specialisations.
Beginning in 2009 with the Financial Reform Act (Subtitle C of Title IX), however, the USA initiated a campaign to eliminate references to NRSRO ratings in certain statutes. The 2010 Dodd-Frank Act (Section 6009) continued this expungement. Reducing the mechanistic reliance on CRA ratings is also advocated by the Financial Stability Board (FSB).4 Yes, alternative provisions first have to be identified and implemented. But an EU CRA and ESMA registration could have the opposite effect and institutionalise the significance of external ratings rather than remove them. At that stage, viable alternatives would be even more difficult to devise and operationalise. Even if the EU decided to expel Moody’s or S&P, what would prevent them from issuing ratings from their headquarters in New York?
By officially sanctioning the current practice, without correcting the fallacious analytics of ratings, I argue that the EU enhances the legitimacy of external, exogenous forms of assessment. Ratings, per se, are not problematic. Although ratings address the problem of asymmetric information between issuers of debt and investors, their informational value and marginal utility is minimal given that much of this knowledge already has been priced into market expectations. More sophisticated investors (e.g. PIMCO, Paulson & Co.) perform their own internal risk assessments and don’t rely on Moody’s or S&P to help them understand and evaluate creditworthiness. Arguably, a primary appeal of ratings is as an inexpensive form of outsourced due diligence. Failure, however, to conduct proper internal risk assessments often precipitates a crisis. External ratings may represent value of simplicity, but accuracy suffers.
Given the uncertainty in calculating the risk of sovereign default, investors attempt to minimise such costs while searching for potential arbitrage opportunities. Irrespective of their actual quality, as regulatory licences, ratings provide the chance for investors to capitalise on the creditworthiness differentials of Member States. Disparate governments become synchronically connected and comparable as ratings entitle them to varying degrees of accessibility to liquid capital markets. In other words, speculators now have the instruments with which to exploit the relative vulnerability of individual governments. Would these market participants even listen to an EU-sponsored agency making claims about the health of its distressed masters? One cannot help but be incredulous of such assertions.
Without doubt, the constitution of authoritative knowledge relies on credibility. Given the tremendous intersubjective barriers to entry, market share in the ratings space is not easily gained; hence the present hegemony of the Big Two and a Half – Fitch is a distant third. In 2009, Moody’s, S&P and Fitch issued an astonishing 97 per cent of all outstanding ratings across all categories. An EU CRA would need to convince these market participants – essentially the entire ratings market – to abandon Moody’s or S&P and pay it to assess their creditworthiness. A solid reputation for impartiality and competence would be essential for its success. Here the EU’s patronage can diminish any credibility which this quasi-public agency seeks to establish. Thus, an EU CRA must possess a substantial degree of independence. Of course, that is easier said than done.
Moody’s and S&P’s authoritative ascendance dates back to the rise of market surveillance mechanisms in the mid-nineteenth century. Henry V. Poor was one of the first to systematically document the growing American industrial complex with the 1860 publication “History of Railroads and Canals in the United States”. Industrial statistics occupied John Moody’s 1900 “Manual of Industrial and Miscellaneous Securities”, which included information about the stocks and bonds of financial firms and government institutions. Unfortunately, an EU CRA would not have the luxury of such a grace period. With the arrival of new CRAs, whatever minimal market share exists will further diminish, thus forcing many smaller firms out of the industry altogether while elevating the status of Moody’s and S&P.
The current sovereign debt crisis only complicates this credibility dilemma. Although at first adamantly opposed by the European Central Bank, after their 21 July 2011 Euro Area Summit, EU politicians finally admitted what financial markets had long suspected: Greece has little alternative but to restructure its debt obligations. Reiterating their previous warnings, the main credit rating agencies announced that Greece’s failure to meet its interest or principal payments in a timely fashion or on “less favourable terms” constitutes “selective” default.5 Moody’s confirmed that “the probability of a distressed exchange, and hence a default, on Greek government bonds is virtually 100%”.6 On the one hand, if the real objective of an EU CRA would be to lessen the burden on beleaguered Member States then it would be a farce. By no means is an EU CRA a viable solution to the chronic uncompetitiveness and fiscal profligacy that plagues the periphery economies of EMU. On the other hand, if this agency were to possess all the capacities and independence of a Moody’s or S&P then it would be exactly like them. Why would it behave in any other way? Again, both approaches would simply compound the difficulties facing the EU in managing effectively its sovereign debt crisis.
Conflicts of Interest
Discussions of potential conflicts of interest often focus on the remunerations models employed by rating agencies. As profit-maximising entities, CRAs earn the vast majority of their revenue from the fees that they charge issuers of debt. But this was not always the case. Prior to the creation of the NRSRO designation in the 1970s, subscription fees were the norm. Today, however, about 90 per cent of CRA income is derived from user fees. McGraw-Hill Financial, the parent company of S&P Ratings, generated a revenue stream of US$2.9 billion in 2010, while rating in excess of US$32 trillion in outstanding debt. With reported revenue of US$2.03 billion in 2010, Moody’s Corporation was also quite profitable. Its first quarter revenue for 2011 jumped 21 per cent to US$577.1 million. Both Moody’s and S&P have similar fees. S&P charges issuers of corporate debt up to 4.25 basis points for most transactions, with a minimum fee of US$70 000. Sovereigns can pay anywhere from US$60 000 to US$100 000. Although ancillary practices, such as consulting, contributed to their business, stipulations in the EU regulatory framework prohibit CRAs from providing advisory services.
Revenue dependence on user fees has been suggested as a potential source of conflict of interest. Grade inflation may attract more clients and thus greater profits, as issuers have an incentive to “shop around” as they solicit the most favourable assessment. Such concerns may be warranted in regards to smaller CRAs. Moody’s and S&P, however, are so well entrenched and command such tremendous market share that they are virtually immune from such pressures. Moreover, given the fragility of credibility, rating agencies have a vested interest in preserving and enhancing their reputation for impartiality. Although an EU CRA may not be primarily motivated by the bottom line, grade inflation is a tempting low-cost and highly effective strategy to entice new clients away from its well established competitors. Of course, in its nascent stages of trial and error, it is completely reasonable to expect a degree of overzealousness and inconsistency. Whether issuers would tolerate this volatility given the available alternatives is doubtful.
In addition to the “user-pays” model, in its 2010 consultation paper, the Commission identified other possible options to stimulate competition between CRAs. The “subscriber/investor-pays” model would require institutional investors to obtain their own rating before they can buy a financial instrument. Issuers of debt would select the rating agency of their choice. The ambition is the creation of a “subscriber-pays” rating market. However, the “issuer-pays” approach was a response to the “free-riding problem” of non-subscribers accessing published ratings. In today’s information society, the confidentiality upon which the “subscriber-pays” model rests would be impossible. Furthermore, the unsolicited ratings which this standard promotes can be deployed in a coercive fashion in order to increase a rating’s circulation. Moody’s conflict with the German reinsurer Hannover Rückversicherung AG is indicative of this dark side to unsolicited ratings. What would prevent an EU CRA from engaging in such tactics in order to drum up business? As it stands, the EU remains ambiguous about which business model, if any, it should endorse.
As troubling as the above issues are, where a conflict of interest is most blatant is in having an EU-sponsored rating agency assess the creditworthiness of its sovereign masters. Rating one’s own debt is laughable. It is doubtful whether Brussels is actually concerned with how well Moody’s or S&P appraise the economic health of a country such as Pakistan or Chile. The Commission is witnessing the disintegration of the eurozone and it wants a countercyclical safeguard to protect its ailing economies from any future onslaught. Whether there is actual merit to this position is really irrelevant. Market optics would portray a quasi-public EU CRA as a puppet of the national governments with which it is affiliated. This would destroy its credibility and tarnish the reputation of the Member States.
Two-Tier Rating System
Few would dispute that increased competition in the ratings space would be positive. Granting rating agencies access to the information of issuers which do not employ them may promote independent ratings. But even the ECB questions if an EU CRA would enhance competition or simply erect artificial barriers to entry to the detriment of private rivals.7 Taking into consideration the arguments outlined above, it is difficult to imagine how a quasi-public EU rating agency can accumulate the necessary reputational capital to compete with the likes of Moody’s or S&P. If their oligopoly can be disturbed, then it will not be instantaneous. In fact, as I contend, an EU CRA may have the adverse effect of further cementing their dominance. This would create a two-tier rating system where the EU CRA plays a peripheral role relative to Moody’s or S&P. To level the playing field and compel market participants to adopt the new agency’s ratings, market distorting mechanisms would be necessary. Otherwise, this asymmetry would undermine the position of the new EU CRA as a reputable alternative.
Preoccupation with the quantity of ratings, however, distracts attention away from the real problem: their dubious quality. Rating agencies have a dismal record of predicting the deterioration of economic positions. Only in less than 25 per cent of cases have Moody’s and S&P cut a sovereign rating before the onset of a correction. Most downgrades occur once a crisis has already begun. Risk calculus divorces ratings from the messy and uncertain world of fiscal politics. It imposes an artificial budgetary normality onto the European political economy, thereby attempting to eliminate the alterity that exists between Member States.
Heterogeneous economies, however, cannot be reduced to probabilistic estimation of risk default. Uncertainty cannot be calculated and (mis)represented as a risk. CRAs claim not to design ratings as a probabilistically quantifiable frequency denoting the credit event of default or expected loss but rather ordinal rankings of credit risk. Nevertheless, key (qualitative) political determinants, such as the stability and legitimacy of political institutions or the transparency of policy decisions, are framed in absolute risk terms in order to be tractable to the rational choice scenarios and stress tests implicit in CRA propriety models. Without any clear alternative to measuring creditworthiness, an EU CRA is bound to adopt this fallacious analytics of ratings. Thus, the end result will be even more suspect external assessments that threaten to cause even more instability and undermine EU efforts to manage its sovereign debt woes.
There is no simple method to regulate and sanction informal judgement. Although additional measures are necessary to compensate for the inadequacies of the existing CRA framework, I posit that a quasi-public EU rating agency is not the solution. Rather than reducing the mechanistic overreliance on external forms of assessment, an EU CRA can have the effect of actually heightening this dependence. ESMA registration can contribute to institutionalising the status of ratings while it dilutes the lower rungs of the ratings industry. That can only enhance the legitimacy of Moody’s or S&P. Severe conflicts of interest will compound these challenges. Neither does this proposal redress the fallacious analytics of ratings and their poor quality. Overall, an EU CRA can undermine the EU’s authority to manage effectively the sovereign debt crisis.
- 1 European Securities Markets Experts Group: Role of Credit Rating Agencies, Report to the European Commission, Brussels, 6 June 2008.
- 2 Commission of the European Communities: Public Consultation on Credit Rating Agencies, IP/10/1471, Brussels, 5 November 2010.
- 3 Moody’s Investor Services: Moody’s Rating Methodology: Sovereign Bond Rating, New York 2008, p. 4.
- 4 Financial Stability Board: Principles for Reducing Reliance on CRA Ratings, Basel 2010.
- 5 Standard & Poor’s: When Would A “Reprofiling” Of Sovereign Debt Constitute A Default?, New York 2011.
- 6 Moody’s Investor Services: Moody’s Downgrades Greece to Ca from Caa1, Developing Outlook, New York 2011.
- 7 European Central Bank: European Commission’s Public Consultation on Credit Rating Agencies – Eurosystem Reply, Frankfurt 2011.
Contributors to this Forum
Gunther Tichy, Austrian Academy of Sciences, Vienna, Austria.
Karel Lannoo, Centre for European Policy Studies, Brussels, Belgium.
Owain ap Gwilym, Bangor University, Bangor, UK.
Rasha Alsakka, Bangor University, Bangor, UK.
Donato Masciandaro, Department of Economics and Paolo Baffi Center, Bocconi University, and SUERF, Italy.
Bartholomew Paudyn, University of Victoria, Canada, and University of Warwick, UK.