How the XVA/CVA trading desk of a leading Canadian bank uses Credit Consensus Ratings before the trade goes on the books.
The Problem
XVA pricing depends on the desk’s view of counterparty credit. Liquid names answer that directly — CDS curves, traded bonds. But the majority of derivative counterparties have no observable credit spread, so the desk prices off internal ratings and a proxy-mapping methodology. Three things make that view expensive to get wrong:
Mispricing is locked in
Once an uncollateralized trade is on the books there is typically no liquid CDS to hedge it and no recourse to reprice the client — for the life of the trade.
Capital follows the rating
Under SA-CVA, regulatory capital is driven by the counterparty rating — especially on proxy names where no other credit view exists.
Stale views select against you
A desk that is slow to downgrade wins exactly the trades it has underpriced — and watches the credit deteriorate after they are on the books.
How one desk uses consensus data
| 1 |
Early warning before adding exposure — the desk's top priority Before putting on material trades, the desk checks the direction of the consensus. A rising consensus PD against a static internal rating is a signal to pause and ask: can we win this trade — do we want to? The desk's stated top concern is being too slow to downgrade; the consensus is its independent trend check. |
| 2 |
Pre-trade pricing calibration The desk compares its internal rating to the consensus on each counterparty. On most names the internal view sits within half to one notch of consensus — which the desk treats as validation of its process. The actionable signals are the outliers of a notch and a half or more: price too wide and you lose the trade; too tight and you systematically win the trades you have underpriced. |
| 3 |
Two-track pricing with sales Front-office pricing can flex commercially trade by trade, while end-of-day valuation and reserving stay tied to the internal rating. Derivative sales see both the internal rating and the consensus, and desk and sales decide together whether to sharpen pricing on a specific trade. |
The pricing asymmetry
Win the trades the market says you are pricing too wide. Avoid the ones you would win only because you are pricing too tight.
The desk’s quant group automated a bulk download of the data, which also feeds a peer-comparison report — internal ratings versus consensus across the bank’s submitted universe. Some desks are also exploring the consensus as the external source rating within their approved proxy-mapping methodology; an internal risk-oversight conversation, not established practice.
What the data is
Credit Consensus Ratings aggregate the internal credit views of a consortium of IRB banks into entity-level consensus ratings (21-notch CB21 scale mapped to PDs), with consensus depth, rating trends, sector and geography PD indices, and transition matrices. Coverage concentrates where public signals are scarce: no CDS curve, no traded bonds, no agency rating. Delivered via web application, bulk feed, and Bloomberg.
Next step
Request a peer-comparison analysis on your counterparty universe — your internal ratings against the consensus, scoped to a defined counterparty list or peer group.
Contact [email protected] to request your analysis.