In credit assessment, alternative data can refer to any type of data that is not considered in the traditional bureau score or considered as demographic data.
Common alternative data sources in credit scoring include device data, psychometrics, open banking, digital footprint or transactional data.
AUC (ROC Curve)
AUC stands for Area Under Curve, in the context of data science it is used as a performance metric of how well a model can identify two different groups.
An AUC is shown as a measurement between 0 and 1, however, in the context of credit scoring, it is expected to observe results between 0.23 and 0.7 depending on the data available.
In the context of credit assessment, behavioural insights evaluate a set of character or behavioural traits which have been shown to be predictive of someone’s likelihood to repay or default.
Some examples of behavioural insights include: Ambition, Entrepreneur, Optimism, Time Management etc.
There are various ways to gain insights into behaviour, including assessing device data or psychometric assessment.
A ‘Credit Invisible’ may be used to refer to an individual with no formal credit history, making them invisible to financial institutions relying on traditional credit scores for assessment. Without a credit history, the traditional methods are unable to ‘see’ these individuals.
A credit score is an analysis of an individual’s creditworthiness, this is generally represented by a three-digit number.
Traditional bureaus scores established the practice of giving a number between 150– 950, with a higher number indicating the applicant is more worthy of credit.
Recently, more institutions are realising that traditional scores take into account a limited set of data based only around an individual’s ability to repay. Alternative Data credit scores consider character and behaviour to give an indication of the individual’s willingness to repay.
Similar to digital footprint, device data refers to traceable data from your device – normally a mobile phone.
This data is accessed via an app after the user has granted permissions. Begini does not access any private information. We look at the way you use your device, an example may be, how often the phone is recharged.
A digital footprint refers to traceable data that is linked to an individual. This could be online or on any digital device. It may refer to actions, communications, or contributions.
Every individual has a unique digital footprint.
Equal Credit Opportunity Act
Begini follows the Equal Credit Opportunity Act which disallowed credit-score systems from using information like sex, race, marital status, national origin and religion within the credit scoring models. Therefore, even if we unknowingly collect such information from customer’s devices, we will not use it within our models.
The General Data Protection Regulation (GDPR) is a regulation in EU law on data protection and privacy. It is currently considered the strictest data protection regulation on the planet.
The Gini Coefficient (or Gini Index) is an industry term for evaluating the predictive power of a credit risk model.
It is an indication of the models ability to discriminate between a ‘good’ or ‘bad’ borrower.
The Gini is measured on a scale of 0 to 1, with a higher Gini being more predictive. In the credit industry, it is usual to observe Ginis between 0.2 and 0.7.
A Gini will change over time and should not be seen as ‘fixed’. A credit model should be periodically tuned with new repayment information, which will improve the Gini.
Model (in regards to data science)
A data model is a way to organize elements data to explore how they relate to each other. In the context of credit scores, data models can explore how elements of data relate to outcomes, specifically the likelihood of default.
Personally Identifiable Information (PII) is any data that could potentially identify a specific individual
Privacy & Regulations
Each region has their own data privacy regulations. Begini does not collect PII and complies with all GDPR regulations (currently the strictest regulations on the planet).
Psychometric assessment is a scientific approach, within the field of psychology, used to measure an individual’s character and/or capabilities.
Psychometric Assessments have been used in various domains for many years. Within the context of credit assessment, psychometric tests can be used to understand an individual’s willingness to repay.
In regards to credit assessment, a thin-file applicant is an individual with little or no credit history, meaning they do not have enough information in their file to be assessed using traditional credit assessment (i.e. bureau scores).
For ‘Thin File’ customers, alternative data provides a way to be financially included.
Willingness to repay
While traditional credit assessment consider an applicant’s ability to repay by looking at historical repayment information, an alternative credit score will indicate an applicant’s willingness to repay by considering character, behaviour, skills and abilities.
If you have a suggestion to add to this list, or would like a term defined by our experts, send it through on our Contact page and we will follow up.