Economic policy uncertainty and sovereign credit rating decisions: Panel quantile evidence for the Eurozone

http://dx.doi.org/10.1016/j.jimonfin.2017.08.007 0261-5606/ 2017 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ⇑ Corresponding author at: Management School, University of Liverpool, Liverpool L69 7ZH, United Kingdom. E-mail addresses: P.Boumparis@liverpool.ac.uk (P. Boumparis), costas.milas@liverpool.ac.uk (C. Milas), tpanag@uom.gr (T. Panagiotidis). 1 For instance, the International Monetary Fund estimates that gross debt in thirty-nine advanced economies deteriorated from 71.2% of GDP in 107.5% in 2016 whereas gross debt in the Euro area deteriorated from 64.9% of GDP in 2007 to 91.7% of GDP in 2016. Data available from: https://www external/pubs/ft/weo/2016/02/weodata/weoselagr.aspx. 2 In particular, Greece, Ireland, Italy, Portugal, Spain and Cyprus have been downgraded by 14, 10, 6, 10, 9 and 14 notches, respectively by Mood 3 The three main CRAs have a total EU market share of 92.85% (see https://www.esma.europa.eu/sites/default/files/library/20161662_cra share_calculation.pdf). Periklis Boumparis , Costas Milas a,b,⇑, Theodore Panagiotidis c


Introduction
During the global financial crisis of 2007-2009 and the subsequent recession, Central Banks and governments responded by injecting additional liquidity into the system and pursuing expansionary fiscal policies, respectively. With the world economy in (the process of returning to) normality, fiscal positions are also being tightened up. Nevertheless, the significant deterioration of public finances post 2007 1 has put on alert Credit Rating Agencies (hereafter CRAs). For instance, Moody's Investor Services, a major credit rating agency, has downgraded over the 2008-2013 period the debt rating of a number of peripheral European countries, namely Greece, Ireland, Italy, Portugal, and Spain (hereafter the GIIPS) and Cyprus by 63 notches in total. 2  Sovereign credit ratings provide a measure of the probability that a country will default on its debt obligations. In that sense, they set the tone for gauging borrowing costs in international markets both for a sovereign state and the financial institutions operating in that sovereign state (for recent evidence, see Drago and Gallo, 2017). This is vital for stimulating investments and supporting economic growth.
Reputational concerns do discipline the decisions made by CRAs (see e.g. Bar-Isaac andShapiro, 2013 andMariano, 2012). However, the value of reputation depends on economic fundamentals that vary over the business cycle. Using a theoretical model of credit ratings with endogenous reputation, Bar-Isaac and Shapiro (2013) relate credit ratings decisions to the economic cycle. They find that CRAs are more likely to issue less accurate ratings when fee-income is high, the economy is booming and securities' default probabilities are low. Indeed, during booms, hiring skilled analysts becomes more expensive for CRAs. At the same time, CRAs can potentially charge higher fees and since bond issues are less likely to default, monitoring a CRA activity becomes less effective.
Although the recent empirical literature has discussed a number of quantitative and qualitative factors affecting the decisions of CRAs, an increasingly large number of decisions appear to remain unexplained. For instance, some of the downgrades of peripheral European debt which took place in 2010 and beyond have been contested by the downgraded peripheral countries and by prominent European policymakers. Speaking to the European parliament in May 2010, Jose Manuel Barroso, then the European Union Commission President, criticised the three main CRAs noting that ''deficiencies in their working methods has led to ratings being too cyclical, too reliant on the general market mood rather than on fundamentalsregardless of whether market mood is too optimistic or too pessimistic" (Barroso, 2010).
In a letter published in March 2011 by The Economist, David Beers (2011), Standard & Poor's (at that time) Global head of sovereign ratings, defended the record of the CRAs. He noted that credit ratings ''provide a robust ranking of the risk of sovereign default" and ''are independent opinions of creditworthiness based on fundamental analysis and therefore should be expected to change as credit risk evolves over the cycle". Gärtner and Griesbach (2012) argued that ''sovereign ratings, their meaning and their underlying procedures are rather opaque". They also went on to argue that ''the set of relevant fundamental variables is an open one, and the interpretation of ever evolving political institutions and processes in unprecedented environments are a dime a dozen". Moritz Kraemer, Global Chief Rating Officer of Standard & Poor's, dismissed the arguments of Gärtner and Griesbach (2012) as ''simply wrong" and went on to note that S&P's sovereign rating decisions are accompanied by comprehensive published rationales and, often, press releases that explain their reasoning and approach. Kraemer (2012) also pointed out that S&P's explain on their website how they arrive at their ratings and how their ratings perform over time (see www.understandingratings.com) which makes their publications as transparent and complete as possible.
The growing dissatisfaction across Europe about some of the recent credit rating decisions, has given rise to talks among Eurozone member states about setting up a European credit rating agency which will increase competition in the rating business. Nevertheless, the European Central Bank (ECB) has been very cautious about how quickly such a project could be deployed. In February 2011, the ECB pointed out that a new credit rating agency will have to rely on extensive data, a number of models, experienced staff and go through building a sound track record for several years before it establishes itself as a credible agency in the rating business (Tait, 2011). In 2016, European Securities and Markets Authority (ESMA), which is the authority competent for the supervision of CRAs, published a report on sovereign ratings processes which noted that because of a ''switch to a regulated industry with focus on integrity of process . . . ESMA has driven significant changes in the credit rating process and the methodology . . . thereby strengthening their integrity, independence, quality and transparency" (ESMA, 2016 Report, p. 16).
This paper attempts a comprehensive assessment of credit rating decisions made by the three main CRAs for the Eurozone economies in light of the ongoing criticism discussed above. The existing literature on the determinants of sovereign credit ratings has focussed on several macroeconomic, qualitative and risk factors. Recent studies focus on time-varying models of credit ratings (Reusens and Croux, 2017) and models with debt levels conditional on debt being above or below endogenously determined debt threshold levels (Hmiden and Cheikh, 2016). Prior to this, Afonso et al. (2011) examine differentiations across rating levels by splitting their dataset into two groups according to the ratings level, namely high-rated countries with credit grades BBB+ and above and low rated countries with credit grades BBB and below.
Arguably, however, the actual degree of importance of the different explanatory variables across the conditional distribution of sovereign credit rating has not been explored in detail as most of the studies focus on the average responses.
We fill the gap in the literature by implementing panel quantile estimation with nonadditive fixed effects as proposed by Powell (2016). Our contribution to the existing literature is summarised as follows: First, we employ a panel quantile framework that allows us to observe the relative importance of quantitative and qualitative factors across the conditional distribution of sovereign credit ratings rather than merely focussing on their conditional mean. Second, we augment the information set considered in previous studies by examining and identifying the significant impact of competitiveness and the European economic policy uncertainty index on the Eurozone sovereign credit ratings.
Among our findings, the unemployment rate, regulatory quality and competitiveness have a stronger impact for low rated countries whereas GDP per capita is a major driver of high rated countries. A reduction in the current account deficit or an increase in the current account surplus leads to a rating or outlook upgrade for low rated countries. Economic policy uncertainty impacts negatively on credit ratings across the conditional distribution; however, the impact is stronger on the lower rated countries. We quantify the effects of uncertainty on credit ratings by using estimates of our model under uncertainty to infer what credit ratings would have been had uncertainty remained at its pre-financial and pre-European debt crisis average