Country Risk: CDS spreads fail to signal increased risk as quickly as ECR’s survey

Although both indicators provide useful measures of sovereign risk, the superior information content contained in ECR score trends seemingly provides a more important early-warning indicator for bond investors, Jeremy Weltman reports.

A recent analysis by Euromoney Country Risk sought to discover whether the spreads on CDSs were becoming less reliable as an indicator of market risk in light of recent changes in regulation aimed at clamping down on naked short-selling.

The conclusion was that they probably were, and that the regulatory change was at least partly attributable to the tightening of spreads witnessed between July and November of this year, before the change took effect.

This follow-up analysis goes further and seeks to investigate whether CDS spreads were as quick to signal the increased risk associated with the European debt crisis during the past couple of years, or whether economists and other country-risk experts taking part in Euromoney’s Country Risk Survey were more perceptive. A rigorous statistical approach might well answer the question more scientifically, but the use of charts still proves highly illuminating.

The heavily indebted periphery

Greece is excluded from the analysis due to some gaps in the CDS data – the series starts in October 2010 and comes to a halt late in 2011 due to the default; ECR has an unbroken series of data.

However, the other four peripheral eurozone indebted sovereigns – Italy, Spain, Ireland and Portugal – are plotted below. In each case, the CDS spreads (shown on the left-hand axis) are plotted against the inverse of the ECR scores (on the right-hand axis). This ensures that rising lines – equivalent to an increased (loosening) spread and a lower ECR score – move in tandem, symbolizing higher risk. The converse is true whenever the lines fall.

The CDS spreads are averaged from daily data obtained from Markit, the financial information services company, and are representative of the cost of insuring against a sovereign default. A CDS spread of 200 basis points, for example, implies it would cost $200,000 per annum to insure $10 million-worth of debt over a particular time frame, in this case five years.

ECR scores, on the other hand, represent a weighted average of values given to 15 of the most important political, economic and structural risk factors in an online survey of experts. These are combined with values for access to capital, credit ratings and debt indicators, to give an overall score – and one that could be said to provide a more complete picture of sovereign risk than simply looking at the constituent aspects alone.

The four indebted-periphery charts, shown above and below, either appear to favour the ECR survey or, at the very least, suggest there is no additional information value (signalling content) in CDS spreads. For two of the sovereigns, Ireland and Portugal, ECR scores and CDS spreads appear to move in sync, providing similar information about the change in risk.

The same result was found in the case of Cyprus, now on the brink of default, and in the partial data for Greece.

However, for two more, Italy and Spain, the trend shift in risk witnessed in the ECR survey does appear to have occurred earlier than market movements in the CDS market. Incidentally, this was also the case when the ECR survey was compared with credit ratings actions by the three main agencies – Fitch, Moody’s and Standard & Poor’s ECR scores responded more quickly than the credit ratings.

Note that in the case of Italy (below) – ignoring the recent divergence, which was explained in the previous analysis – the ECR score fell between September 2010 and March 2011 as ECR experts became jittery about default risk, a trend that continued in subsequent months.

However, the CDS spread was still falling (tightening) between December 2010 and April 2011, signalling a lower insurance required against such a default, and therefore a lower level of risk.

Of course, risk is a flexible concept – Italy has not defaulted, but the threat of default has increased as the sovereign’s ability to meet its fiscal obligations has become more difficult and this is manifest in a range of market indicators, including elevated borrowing costs (bond yields) and lower credit ratings.

This temporary tightening – which, with hindsight, appears to be the inefficient signalling of risk – seen in CDS spreads is also true of Spain (below). Here, the ECR score shift, denoting increased risk perceptions among experts, similarly preceded the CDS market spread reversal.

Note the lower ECR scores from September 2010 onwards, but tighter CDS spreads between December 2010 and April 2011. Taking out annual insurance on $10 million-worth of Spanish sovereign debt over five years would have cost approximately $100,000 less in April 2011 than December 2010, yet Spain’s fiscal metrics were still acute. Its general government deficit exceeded 9% of GDP in both 2010 and 2011, and its debt trajectory showed an incline away from the golden limit of 60% of GDP. This is before the weak economy and banking system woes were factored in.

The ECR scores were therefore providing an earlier – and it could also be said a more efficient – signalling system to the risk of debt-default, which has been now well-documented for both countries during the past couple of years.

Further east

Two of the central and eastern European sovereigns exhibiting a distinct trend-rise in risk (lower ECR scores) during the past few years are Hungary and Slovenia (plotted below). The problems affecting both countries have been regularly reviewed by ECR (for Hungary see: Hungary still riskier than its CDS trend suggests and ECR score change: Hungary; and for Slovenia: Seven CEE borrowers have riskier government finances than Italy; Slovenia plunging fastest and Slovenia leads rise in bank-stability risk across Europe).

CDS spreads, however, follow a similar pattern to the previously mentioned sovereigns. The discrepancy is pronounced in Hungary’s case, with an average decline (tightening) in the spread of around 138bp from December 2010 to April 2011.

The Slovenia CDS spread also tightened, though less markedly, despite a widening general government deficit – to 6.4% of GDP in 2011 from 5.7% of GDP in 2010 – a rising level of debt, weak GDP growth and banking system troubles.

Even among the top-rated sovereigns

The UK has not escaped either, with three distinct patterns of CDS spread tightening. The first occurred to October 2010, the second was the drop from December 2010 to April 2011 familiar to other countries, and a third took place from December 2011 to March. This was subsequently partially reversed before the drop since mid-2012 also seen in other countries and referred to in the previous analysis.

The UK’s ECR score, however, has followed a trend decline during the entire period and its Aaa credit rating has been placed on review for a downgrade.

Further information on the survey is available from Euromoney Country Risk.