When financial markets participants are bored or dyspeptic, they frequently turn to credit rating agencies as an always-available target of criticism.
The three main rating agencies, Standard & Poor’s, Moody’s and Fitch, have been seen as at least partially culpable for elements of the financial crisis, from the fiasco of subprime mortgage securitisation to tipping Greece into disaster when its sovereign credit rating was downgraded.
Their actual ratings have also come under attack, with academic papers and bank economists analysing them to find evidence for home bias, subjective error and over-lenient analysis.
Their business model is doing just fine, however, particularly in the bond ratings business. The issuer-pays model, much criticised during and after the crisis, remains in place, while the embedded regulatory support for the rating agencies that have Nationally Recognized Statistical Ratings Organization status in the US continues to underpin the structure. Of the big three, profits at S&P and Moody’s are at or near record highs.
Now there is at least one serious attempt to create an alternative to the traditional commercial model for sovereign credit ratings.
The Bertelsmann Foundation has mooted the establishment of an international non-profit credit rating agency (Incra), which they propose should be funded as a sustainable endowment. It calls for an initial injection of $400m, to be gathered from a broad range of donors, including governments, supranational bodies such as the World Bank and the International Monetary Fund, civil society in the form of foundations and non-governmental organisations, and the financial services industry.
“If you look around the world, there is a boom in new rating agencies,” says Annette Heuser, executive director of the Bertelsmann Foundation. “This tells us there is demand for them.”
In a 70-page document (available on its website), Bertelsmann outlines the complex governance structure for Incra, which attempts to create confidence in the body’s independence by ringfencing the analytical function from the funders with a stakeholder committee.
It also goes into extensive detail on the methodology (although not the precise formulas) that would be used for calculating ratings. Despite this careful planning, Moritz Kraemer, chief sovereign rating officer at S&P, is dismissive of the idea.
“I don’t think they are ratings in the way that investors need them to be,” he says. “They are more like what the Economist Intelligence Unit puts out.”
What investors need, he explains, is ratings that use the standard terminology using single, double or triple letters (A, B or C) combined with positive or negative signs. Incra would use a simple numerical scale.
“If you use a scale no one is familiar with, no one knows what that means,” says Mr Kraemer. “There is a reason a lot of the agencies have ended up using the same nomenclature.”
He is also dubious about the credibility of a new set of ratings, which will be proved only through back-testing, rather than years of experience.
The downside of having years of published ratings is that they are vulnerable to analysis by outsiders. In 2012, academics at the University of St Gallen in Switzerland found sovereign credit ratings had sufficient impact on countries’ interest rates (the cost of borrowing on capital markets) to generate multiple equilibria. As long as an economy’s rating remained at A or above, all was stable, but as it slipped down, an unstable equilibrium could be reached. In this scenario, the country’s interest rate could be pushed so high that the economy was then pushed further into trouble as it could only borrow at unsustainable rates.
Another study from the University of Heidelberg found an appreciable home bias in sovereign ratings, whereby agencies cut their home country greater slack, particularly after the onset of the financial crisis.
A detailed criticism of the current sovereign ratings methodologies and a proposal for an alternative comes from Erik Nielsen, chief economist at UniCredit, the Italian bank. In a paper published with colleagues Daniel Vernazza and Vasileios Gkionakis, Mr Nielsen contends the output of all three big agencies could be improved by using an algorithm derived from a simple regression of their output, and cutting out the subjective element that is represented by the error term of the regression.
This implies, according to Mr Nielsen, that the element of judgment, the committee-based discussion that is the core of most ratings processes, actually makes the ratings less accurate, rather than more so. We would be better off feeding the numbers into a computer and just accepting the readout, in other words.
Mr Kraemer is dismissive of this suggestion: “The key problem with the UniCredit research is that it pretends it can distinguish between objective and subjective elements of the ratings.”
The UniCredit suggestion of simply developing an algorithm and letting a body such as the World Bank oversee its functioning is simpler than most. Relatively few experts are keen to see personal judgment cast aside for mechanistic calculations, however much effort is put into gathering the data. Incra, for example, sees itself maintaining a large number of region and country bureaus to gather information, but would then revert to the committee-based format to tease out the precise meaning of the data.
Some of Mr Nielsen’s criticism of the major agencies is out of date, according to Mr Kraemer. The UniCredit paper is disapproving of the opacity of rating methodologies, but according to Mr Kraemer, that is a thing of the past.
To complaints that variables are listed without any indication of what weights they are given in the calculations, he says: “It may have been the case there was a cookbook that listed the ingredients without saying how much flour or sugar should be used, but that is no longer so.”
The cookbook now includes precise instructions, he claims. He adds that he is disappointed that critics assume information is not available when it is on the website or would be made available to them on inquiry.
This is a recent change, he admits, pointing to this as an example of how the rating agencies have learnt from recent debates.
Mr Nielsen has one other significant criticism of the current credit ratings system: if it is incorrect, it creates significant distortions of capital allocation. In his opinion, the rating agencies overruled signals from the peripheral eurozone countries at the time of the EU debt crisis, so that when the downgrades did come, huge amounts of capital were withdrawn all at once, “contributing to the depth of the recession ... and untold social hardship”.
“This reduces investors to passive actors, just following the rating agencies,” says Mr Kraemer. He points out that countries tend to be downgraded at a time when the fundamentals might suggest investors should move their money, regardless of ratings. “Confusion of correlation and causality is a very basic econometric trap,” he says.