The Handbook of Credit Risk Management, 2nd Edition

CHAPTER 7Alternative Estimations of Credit Quality

Predicting default is a high‐stakes exercise for any entity taking credit risk. An underestimate means the difference between profit and significant losses. An overestimate means a foregone profitable opportunity. Fundamentally, the assessment of default risk drives the yes/no decision about entering a transaction. If the risk is deemed acceptable, the strength of the counterparty dictates the type of transaction, the exposure, the tenor, and, last but not least, the price.

We see in Chapter 2 that the most common methodology for summarizing credit quality is the use of a rating based on an internal review, a rating agency, or a credit‐scoring specialist. Then, a probability of default can be mapped to the rating using the historical default data compiled by rating agencies like S&P, Fitch, or Moody's. This way of estimating a probability of default is the oldest and the most widespread on a global basis. What has changed over time, though, is that methodologies used have become more sophisticated, and data are more plentiful (e.g., historical, peers).

In parallel, the development of quantitative research in the past few decades and the advent of analytics have opened new possibilities. Since the early 1990s, considerable resources have been devoted by researchers and companies of all kinds, from start‐ups to large banks, to develop alternatives to human‐judgment‐driven ratings.

In this chapter, we describe alternatives to ...