Bank reduces risk exposure, boosts return on capital
How a large private sector bank in India built an Early Warning System (EWS) to gauge and assign risk score.

Client
Banks operate in a high‑risk environment and require solutions that can proactively detect and prevent fraud, manage risks, and ensure secure and efficient banking operations. Early Warning Systems (EWS) have proven helpful in this case. Our client, a large private sector bank in India, wanted to build an Early Warning System (EWS) for its corporate clients that could automatically gauge and assign a risk score, probability of default, and EWS rating.
Challenge
Historic month-end data collection spanning four years
Need to identify historic cases that were considered ‘bad’ due to significant weaknesses observed
Need to compute Max Customer DPD to exclude specific products from the DPD computation
Solution
Created a statistical EWS scorecard using a generalized linear model
Identified EWIs using power analysis, correlation analysis, and economic logic
Identified six variables in the final scorecard covering transformed variables related to delinquency, credits in revolving accounts, overdrafts, check returns, etc.
Enabled scorecard scaling, calibration, bucketing, and exit/watch list thresholds
Enabled the tracking of all key EWIs from the core banking system
Impact
Reduced risk exposures, better management of NPA, and optimized return on capital.
Regularly updated scores as relevant data is readily available.
Improved performance of borrower rating model.