The NKR agency published an estimate of the frequency of bond defaults depending on the national credit rating, considering their number in a three-year period. A similar metric is established in the Basel regulatory documents, other rating agencies rely on it, and for bond portfolios, it can be used to assess the risk profile of investments.
In the study, we took a sample of 5,000 observations of Russian companies and assigned a rating to issuers to which they would correspond at each time point. The smoothed frequency of defaults on Russian ratings turned out to be intuitive. A significant proportion of defaults in the model is observed starting with the “BB” rating (8.85%), after which the increase occurs exponentially. For issues with higher ratings, the default rate is estimated at no more than 5.9% over three years, and for ratings of the AAA-A series – no more than 1.44%.
In reality, the ratio of rating to default may not be so statistically correct: the rating may lag behind the state of affairs in the company, and the “test” calculation of the rating can hardly be compared with the actually assigned rating. Therefore, the real distribution of defaults may also be towards higher ratings.
The result provided by the model does not claim to be extremely accurate, but gives a schematic representation of how great the credit risk is in each of the ratings. Considering an issue with a similar rating, we can simplify this metric as an assessment of the credit risk of these bonds.
Author: Ilya Grigoriev