We agree with the principles as laid down by the EBA in it proposed amendments of the proxy spread framework to determine CVA risk. Particularly, we support its proposal to allow institutions to consider the use of alternative credit quality assessments. We believe this would be a helpful addition or substitute to existing methods of measuring credit risk of counterparties where such counterparties have no “peers at all with observed credit spread data”.
However, we encourage the EBA to consider certain amendments to its proposal to make it more robust and ensure it results in the most accurate and representative measurement of credit risk. Specifically, we note that the EBA referred to the use of “credit spread data” in a number of instances, namely when describing its use in proxy spread methodology and when recommending fundamental analysis in the absence of credit spread data. We recommend that the EBA clarify, consistent with the previously issued final RTS on CVA risk for the determination of proxy spreads, that institutions consider not just CDS spreads but also “spreads of other liquid traded credit risk instruments” in the calculation. On this basis market participants would derive credit spread data from other market-based instruments such as bonds where those are sufficiently liquid and representative. To allow for the use of such measures, institutions should be able to quantify the liquidity of the instruments used and demonstrate they are sufficiently liquid.
We agree with the EBA that, where available, market-based measures of credit risk should take precedence over alternative approaches based on fundamental analysis to get an accurate indication of credit risk. To provide institutions with sufficient clarity, the EBA might consider specifying a hierarchy of methods and data sources used to derive credit spreads as basis for CVA calculations. Specifically, institutions could apply the following hierarchy:
1) Single name CDS spread of the institution;
2) Bond spread of the institution – importantly, such spread will still reflect the idiosyncratic credit risk which and should hence be preferable to a proxy spread approach;
3) CDS sector proxy spreads – derived from traded CDS spreads using the attributes of rating, region and industry as specified in the RTS; (if the sub-sector of the which CVA is being calibrated from is sufficiently liquid and the drivers of business creditworthiness are the same) .
4) Bond sector proxy spreads – derived from bond spreads using the three attributes rating, region and industry (if the sub-sector of the which CVA is being calibrated from is sufficiently liquid and the drivers of business creditworthiness are the same); and
5) Alternative approach based on fundamental credit analysis, which should, however, still make use of market-based inputs to the extent possible.
Importantly, even where no relevant CDS or bond spreads are available for the specific name or sector, any fundamental credit analysis should be coupled with the use of relevant market-based measures for counterparties with comparable creditworthiness and characteristics to derive a representative credit spread that will then be used for the CVA calculation.
We agree with amendment that allows for the possibility to adjust the value of the LGDMKT term of the regulatory formula.