CRIF NM GLocal Scoring

What is a Credit Bureau Score?

  • A credit bureau score is a credit risk assessment tool which analyzes and summarizes, into a single numeric score, all the historical files for each consumer or small business that is reported to the credit bureau.
  • This score represents the odds or probability that that person or company will become a “bad payer” on at least one trade line within a specified period of time.

Picture2

 

The Evolution Process of CB Scoring…traditionally

  • Mature Database: A custom risk ranking scoring solution is delivered. This will allow consumers to be identified as a certain level of credit risk from high to low based on the custom scoring solution. Sufficient history in the credit bureau exists.
  • History Development II : the CB information maturity makes possible the expert scoring model refinement.
  •  History Development I: an expert risk ranking score is delivered. This will allow consumers to be identified as a certain level of credit risk from high to low based on the expert score.
  • Inception An expert risk ranking value is provided based on various risk criteria. This will allow consumers to be identified as a certain level of credit risk from high, medium, or low.

imnage 2

 

Creating Stability

What are the key factors that create a stable credit bureau score?

  • Score developed using globally accepted proven standard statistical techniques.
  • Consistent Data Contribution to credit bureau

– # of Contributors

– Type of exposures contributed

– Frequency of contribution

– History of data

  • Management of data contribution as it grows and/or changes

– Monitor data contribution

– Manage the addition of new data sources

– Control codes and format of data reporting

Consider influences on Credit Bureau Score Performance

What impacts the risk trends of a credit score by country?

– Consumer’s personal behavior towards repayment of debt

– Behavior based attitudes.

– Laws applied by specific country

– Data Reporting

– Product usage

– Economy

Factors to measure in a Credit Bureau Score System.

What are the factors must be managed and measured to create a stable credit bureau scoring solution framework?

  • Risk Trends
  • Product contributions
  • Data availability (history and content)
  • Distributions (exclusions, segments, and score)
  •  Contributors

Perform GLocal Methodological Approach

  •  The design of a dynamic scoring solution integrates the score with the status of the credit bureau at various points in time, the economic conditions, the cultural lending/usage values, market and legality of data reporting. Occurs on-going.
  • This is done by focusing on the culture, market, data, and economy.
  • The key is to manage the all of these factors with the data evolution to provide a scoring system.

What are the core principles for performance results?

Data

The data is the backbone to the development and sustainability of a credit bureau scoring solution from inception.Consumer’s Payment Behavior A consumer may be influenced by laws and culture, but ultimately whether a consumer pays or does not pay tends to behave similar characteristics that are predictive of this behavior.

Consumer’s Payment Behavior

A consumer may be influenced by laws and culture, but ultimately whether a consumer pays or does not pay tends to behave similar characteristics that are predictive of this behavior.

Test methodological approach

  • Sample Design for simulation
Observation Period – =varied by bureau maturity (3, 6, or 12 months).

 Outcome Period – 12 months to align with a global standard.

img3

Performance Definition for simulation

  • Bad

  1. Any 90+DPD tradeline within the 12 month performance period.
  • Indeterminate

  1. Not Bad and at least one tradeline was 60DPD within the performance period.
  • Good

  1. Not Bad, not Indeterminate and all tradelines are less than 60DPD and at least 6 months old
  • Other

  1. Credit Bureau report did not have enough data to define performance

 Attributes used for ranking risk

  • 3, 6 or 12 months of history was utilized to create attributes for testing varying levels of bureau maturity.
  •  Priority of Attributes were varied based on the cultural lending
  1. Usage of credit and creating debt.
  2. Minimal credit usage.
  3. Market offerings (type of credit).
  4. Most predictive of risk

Traditional distribution of predictive attributes in Credit Bureau data

Picture123