Credit models are an integral part of credit sanctioning, pricing, risk measurement, monitoring, determining capital and reserve adequacy etc. We help our clients in development of statistical and heuristics credit models for corporate and retail portfolios including PD, LGD, EAD, RAROC, Economic Capital and Stress Testing Models. We assist in calibration of internal models based on internal default and recovery experience as well as external data. For Low Default Portfolios, statistical methods for calibration are used if no suitable external data is available.
We assist institutions in developing PD models for Corporate, Retail, Retail SME and Financial
counterparties. Our capabilities include heuristic model development as well as statistical model development
using Generalized Linear Models (Logit, Probit etc.), Classification Trees etc.
Features of our PD model development exercise include:
- Coverage of various asset classes such as Large Corporate, SME, Retail SME, FI, Retail, Project Finance etc.
- Multiple methodologies such as expert judgment, statistical and various machine learning techniques
- Parsimony, Tractability and Reliability of models
Workout LGD Models are based on economic recoveries from default cases. Our capabilities include LGD Model development based on Regression Trees and GLM.
We develop drawn CCF models based on on-balance sheet conversion rate of off-balance sheet commitments using entropy measures and classification trees.
Undrawn CCF models are developed based on utilization pattern of committed lines using regression trees.
Our Economic Capital Modelling capabilities include Credit VaR and Expected Shortfall computation using Multi-factor Monte Carlo Simulation and allocation of Economic capital, portfolio loss volatility and Expected Shortfall to individual customer, industries and business units.