Credit Score

From Open Risk Manual


A Credit Score is a quantitative (numerical) indicator of the Credit Quality (creditworthiness) of a borrower. A credit score is typically produced after the review (potentially using algorithms) of an individual's Credit History, financial condition and other characteristics.


A credit score is typically used to inform a Credit Decision process (e.g. to extend credit to a new borrower). Credit Scores are used primarily in the context of individuals (consumers, retail borrowers) as opposed to companies and similar legal entities that borrow funds (e.g. countries), in which case one uses the term Credit Rating. The two concepts are related but have also substantial differences

History of Credit Scores

Historically bank lending decisions have been subjective and (at best) rule-of-thumb based processes. In contrast to actuarial science that advanced significantly earlier (19th century), the introduction of statistical techniques in credit decision processes had to wait for the introduction of the digital computer. The reason is instructive of the nature of credit risk which is a substantially more dynamic than actuarial risks (depends on variable conditions).

In the US, FICO was founded in 1956 as Fair, Isaac and Company by engineer William Fair and mathematician Earl Isaac and started commercializing statistical measures of credit worthiness in the subsequent years.

Widespread adoption of credit scoring occured in the nineties.

How are credit scores used?

  • In the first instance credit scores are used to determine who qualifies for a loan via a Cut-Off Score
  • In Risk Based Pricing, the credit score may determine the Interest Rate, and / or the Credit Limit offered to the client.
  • The credit score may be used to determine customer profitability.
  • Credit scores may feed into calculations of regulatory capital which determines minimum requirements for a bank's own funds (equity capital)

How are credit scores produced

A credit score is typically (but not necessarily) the outcome of well defined procedure and algorithms denoted as Scorecards. In the simplest form, a credit score is the weighted numerical sum of variables that have been shown to predict credit worthiness in a statistically significant manner.

Scoring tends to be used for small exposures (consumer finance) and standardised (or homogenous) risk profiles. Individual analyst input is not cost-efficient at a granular level, which is typical of loans to individuals or SME Lending. Ratings are used for other (higher) risk banking products that involve analyst input and quantitative and other qualitative factors.

Producers of Credit Scores


Lenders, such as Banks and Credit Card companies develop and use internal credit scores. These fall in two broad classes

  • scores used in the context of Risk Acceptance to evaluate the potential risk posed by lending money to new clients
  • scores used in the context of Portfolio Management to evaluate the development of credit profile of a borrower

Commercial Entities

Issues and Challenges

The practice of credit scoring faces potentially significant challenges:

  • Algorithmic Bias
  • Difficulty of application for some customer segments, contributing e.g. to the SME Credit Gap
  • Oversimplification of available information: A principal challenge of credit scores is that a complex information set (about the borrower) must be combined into a single numerical assessment. See e.g., [Quantitative versus Qualitative Risk Factors]]
  • Missing possible interactions: The risk profile of the borrower is not completely independent of the nature of the financial product
  • Oversimplification of produced information: A single numerical assessment is used to decide and manage a complex set of future possibilities. For example a credit score does not capture any information as to how the credit quality of the borrower might improve or deteriorate over the course of time (See e.g. Lifetime PD)
  • A credit score as a single metric does not carry any information about the severity of loss (hence must be complemented by other (product specific ) metrics, See e.g. Loss Given Default
  • The credit score isolates only one behavioural aspect of a borrower, whereas in a business context other aspects (e.g. Competing Risks such as prepayment or utilization of financial products) are both informative and relevant for the customer relationship

See Also



This entry annotates a FIBO Ontology Class. FIBO is a trademark and the FIBO Ontology is copyright of the EDM Council, released under the MIT Open Source License. There is no guarantee that the content of this page will remain aligned with or correctly interprets the concepts covered by the FIBO ontology.