Skip to main content
A notch difference is the numerical gap between two PD estimates once both are mapped to a rating scale. Credit Benchmark uses notch differences in portfolio analytics, peer benchmarking, and time-series tracking — anywhere a relative credit quality comparison is needed. By default we use the 21-category CB scale; see the Rating Scale. The same concept applies using a bank’s internal rating scale — if a mapping is provided, notch differences can be calculated on the bank’s own scale. Notch Difference=f(PDi)−f(PDj)\text{Notch Difference} = f(PD_i) - f(PD_j) Where:
  • PDiPD_i, PDjPD_j — PD estimates from two sources being compared
  • f(PD)f(PD) — rating scale mapping function returning a notch index
  • Positive values indicate PDiPD_i is the stronger credit; negative values indicate the reverse

​
Applications

f(PDb)−f(PDc)f(PD_b) - f(PD_c)How far a contributing bank’s view sits from the market consensus for the same entity.
  • Positive — bank is more optimistic than consensus (lower PD)
  • Negative — bank is more conservative than consensus (higher PD)
f(PDt)−f(PDt−n)f(PD_t) - f(PD_{t-n})How credit quality on an entity has shifted over a chosen period — 1, 3, 6, or 12 months.
  • Positive — credit quality improved (PD decreased)
  • Negative — credit quality deteriorated (PD increased)
f(PDA)−f(PDB)f(PD_A) - f(PD_B)Relative credit quality between two entities — used for peer benchmarking and portfolio analysis.
  • Positive — Entity A has stronger credit quality than Entity B
  • Negative — Entity A has weaker credit quality than Entity B