How does KMV-Moody's EDF differ from the plain Merton model?

Question

AcadiFi · Accepted Answer

Moody's KMV (now Moody's Analytics CreditEdge) extends Merton in three important ways. First, the debt point is not just total liabilities — KMV uses short-term debt plus half of long-term debt plus other liabilities, reflecting that firms rarely default when assets cross total debt because long debt need not be repaid immediately. Second, KMV solves for asset value and volatility using an iterative procedure across rolling equity return windows rather than the single-equation Merton solve. Third and most importantly, KMV does NOT map DD to PD via the normal CDF.

Instead, KMV maintains a proprietary database of 40+ years of default outcomes across millions of firm-years. For each DD bucket, they compute the empirical default frequency observed historically. For example, at DD = 4 the theoretical Merton PD = N(-4) = 0.003%, but the empirical EDF from KMV's database might be closer to 0.4% — more than 100 times higher. This reflects that real asset returns have fat tails and firms can default via distress exchanges before crossing the technical default boundary.

On the FRM exam, know that KMV EDFs are a monotone function of DD but are much more conservative than N(-DD) for high-quality obligors. Also know EDFs are forward-looking, updating daily as equity prices move, which is useful for early warning but makes them noisier than agency ratings.

How does KMV-Moody's EDF differ from the plain Merton model?

Master Part I with our FRM Course

Related Questions

Practice Questions