Scoring Service

The CRD model, which is estimated using a huge SME dataset in CRD, predicts the credit risk of SME borrowers. Credit risk is expressed as the predicted probability of default (in percent) and credit score.

Figure: CRD model as a statistical tool that captures signals from financial statements.

The CRD model is estimated using client data of financial institutions with member status. This model is used to determine the guarantee rates in all public guarantee corporations in Japan. In addition, the model is used by financial institutions to improve their assessment of the creditworthiness of borrowers and for advanced credit risk management.

Content of Scoring Service

1. Browser Scoring Service

Using data from borrowers' financial statements on the CRD Association website, members can immediately check the credit score and probability of default of the borrower.

2. Collective Scoring Service

Based on financial statement data of borrowers sent regularly by CRD members, CRD Association provides estimated credit scores and probabilities of default. This service is used regularly (for example, regular monitoring of client credit risk).

3. Requested Scoring Service

Based on financial statement data of borrowers sent irregularly by CRD members, CRD Association provides the estimated credit score and probability of default.

If members provide data on the borrower' s credit rating calculated by their rating system, CRD Association supports to remodel these credit ratings (which involves a mapping analysis using CRD data). This service offers a matrix of credit ratings from the models of both CRD Association and CRD members.

4. Stand-alone Service

This service offers a Windows software package that includes a credit scoring service and a support service for the accumulation of data.

5. Scoring Subroutine Service

This service provides the logic of the credit scoring system in the CRD model as a subroutine (DLL) applicable for members' computer systems.

Page Top