Perspectives 2019 2020 Public Sector

20 Structural Penalties in Sovereign Credit Ratings for many years thereafter. The second and third are examples of variables that only very ambitious governments might be able to meaningfully influence within the normal life of a political administration. While Citi recognizes that none of the ratings factors are easy to improve, the three variables discussed in this article are selected for their particular unresponsiveness to ordinary political, economic and financial policy. The size penalty is particularly interesting, because relative economic — and geographic — size are broadly stable characteristics of a country except in the most extraordinary circumstances, and the ways in which size impacts credit ratings might not be obvious. Hard-to-Change Ratings Factors To assign sovereign credit ratings, each of the agencies employs a proprietary rating model comprised of quantitative and qualitative components. In all three of the Fitch, Moody’s and S&P models, the quantitative component produces an indicative rating or rating range based upon a set of individually- weighted numeric factors. The factor weights in the models are set by complex, multivariate regressions of the factors against historical observations of sovereign willingness and ability to meet debt obligations. Each agency’s model uses a unique but similar set of factors, the relative weights of which are recalibrated at least annually. The calculation-based indicative score is then subject to qualitative adjustments, in which agency analysts can shift the rating to better reflect non-numeric context, information and judgments. 3 Because of the qualitative component in each agency methodology, the quantitative factors are not determinative, limiting both an issuer’s ability to “game” the process for improved ratings and Citi’s ability to predict exact results. Nevertheless, the quantitative models are a robust guide to likely rating outcomes and reflect the relative importance the agencies place upon different sovereign characteristics. It is therefore notable that all three methodologies include a measure of prior issuer defaults over a long historical horizon, potentially allowing the mistakes of past governments to constrain the rating. In Fitch’s 18-factor model, “Years since default or restructuring event” carries an overall weight of 6.6% in the quantitative model. It is measured from 1980 such that an event in 1979 or earlier returns a zero, while an event in 2019 returns a negative one and would result in a roughly 2.6-notch downgrade — a BBB to BB+, all else being equal. 4 Moody’s current methodology considers “Track record of default” over a 20-year period and assigns a score of zero to negative three, “serial defaulters” being assigned the lowest score. Again holding all else equal, applying a score of zero to one specific Baa3-rated CEEMEA issuer would likely result in the loss of its investment grade rating in the quantitative model. 5 S&P treats prior defaults differently, treating the variable as binary and a binding constraint on the final rating. For sovereigns with any of “significant and sustained arrears on bilateral debt,” “a public discourse that questions the legitimacy of debt contracted by a previous administration,” or “no material policy change since the last default,” S&P places a rating ceiling of BB+. For S&P, past instances of debt restructuring or defaults reflect a poor “debt repayment culture.” Indicators of such a culture correlate with a greater likelihood of re-default in future and, as S&P observes, “history demonstrates that countries can graduate from being serial defaulters, although the path to doing so may be long.” 6 Fitch and Moody’s discuss their treatments of prior restructurings in similar terms, thereby explaining the long look-back period for this factor and emphasizing for sovereigns the importance of avoiding debt restructurings and defaults whenever possible. 3 In fact, qualitative judgements are found throughout the quantitative components in each agency’s model as well. For example, we can consider S&P’s assessment of sovereign “Fiscal Performance and Flexibility.” The scores requires calculation of the average change in net general government debt as a percent of GDP including the current-year estimate and S&P’s two- or three-year forecast estimates. This average figure is then transposed into a score of 1 to 6, such that an average of between <0% and 1% becomes a “1,” an average between 0% and 3% is a “2,” and so on until an average <6% is a “6.” Note, however, that an average of 1.9% could be either a “1” or a “2,” in which case the “assessment is decided based on the trend of the government’s fiscal performance.” The space for interpretation is evident — but that is not to say it is problematic. S&P (2017), “Criteria: Sovereign Rating Methodology,” 19. 4 Fitch (2019), “Sovereign Rating Criteria, Master Criteria,” 12. The average factor weight in the model is 5.56%, and the highest-weighted factor is “Governance indicators” at 19.8%. This is calculated as the average of the country’s most recent World Bank Worldwide Governance Indicators scores. 5 Moody’s (2018), “Rating Methodology: Sovereign Bond Ratings,” 16. 6 S&P (2017) “Criteria: Sovereign Rating Methodology,” 9, 16.