This week, in the final instalment of my series on credit ratings, I have a few more thoughts on how to get the most out of ratings subscriptions.
Comparing Ratings Agencies
When comparing ratings across agencies, an effective method is to implement equivalence tables. This is somewhat subject to one’s intended usage of ratings. Also, less commonly used agencies may not publicize suggested equivalencies, which means you or your data management platform provider may have to craft the appropriate equivalences.
One great value in using these tables is that they enable the automated spotting of outliers, the reasons for which could either be a data error or an early signal. The tables also give you the ability to determine composite ratings according to an in-house algorithm. (This translation is slightly trickier for short-term than it is for long-term ratings, as short-term ratings have a more vendor-specific methodology and scale.)
Data Completeness
Suppose you need to calculate the average short-term rating for all securities of a portfolio. In reality, not every security will necessarily have a short-term rating from the same agency. If you’re fortunate, all securities in your portfolio may have at least one short-term rating, though possibly from different agencies. For instance, one security might be rated by S&P and Fitch, while others may have ratings from Moody’s or DBRS. And some may have none at all.
If your data management solution is equipped with potent cross-referencing and backfilling features, you will be able to handle these inevitable data shortcomings well.
Most rating agencies offer a challenge mechanism. Since the resulting data corrections sometimes will have a lag until at least the next update, it is important that the mastering platform supports locking features. This allows preliminary manual updates to not get overwritten by data that is still incorrect in the interim.
That’s a wrap for credit ratings, but please reach out if you have any specific questions – I’m always happy to share what we have learned.