The Data Analytics Blog

Our news and views relating to Data Analytics, Big Data, Machine Learning, and the world of Credit.

All Posts

What is a credit scorecard?

May 19, 2021 at 2:04 PM

A scorecard is a mathematical model that is used to predict a certain outcome. In credit this might be the probability of default. The information used in a scorecard can vary, but common fields include demographic characteristics (e.g. age of applicant, number of dependants, time spent in current job) and credit bureau data (e.g. number of personal loans registered to applicant, worst arrears status on all accounts in the last 6 months).

Scorecards are used throughout developed and many developing credit markets to assist risk managers determine with which customers to do business.

Scorecards are made up of a number of characteristics, typically 7-15. These characteristics represent “questions”. The answers to these questions are known as attributes. For example “How old are you?” would be the characteristic and “28” might be the attribute. The weighting is the “score” assigned to the attribute. Typically the higher the score the lower the risk.

A simple example of a 2 characteristics scorecard is displayed below:

What is a scorecard 1

There are 2 characteristics that appear in this simplified scorecard: “Age of applicant” and “Number of dependants”. Each applicant will be grouped into an attribute group under each characteristic and each attribute will be assigned a weighting. For example Lindsay Loan is 48 (so she gets assigned under the 3rd attribute group for age, namely “46+”). She has 3 children so she falls under the 1-3 dependant attribute group. Next to each attribute group she receives the scores 30 and 15 respectively. Her total score is therefore 30+15=45.

Scorecards typically have cut-offs. Above the cut-off would represent acceptable risk and below the cut-off unacceptable risk. If the cut-off in our above example is 33 then Lindsay would be approved.

Let’s take another 2 applicants.

What is a scorecard 2

The second applicant, Johnny Debt would get 10 points for “Age” and 15 points for "Dependants" again resulting in a total score of 25. Johnny would be declined by the scorecard. The third applicant William would receive scores of 20 and 25 totalling 45 points. William would be accepted. The scores could be represented on the following grid with the coloured squares representing the respective scores of the applicants in the examples (the red numbers being the scores below the cut-off):

What is a scorecard 3

Different Scorecards

Scorecards are used to predict a variety of different outcomes including:

  1. A new applicant defaulting
  2. An existing loan/facility defaulting
  3. A customer in arrears making a successful payment
  4. A customer spending more on her credit card in the next 6 months
  5. A consumer responding to a targeted marketing campaign
  6. A customer lapsing on an insurance premium
  7. A customer closing their account
  8. A customer skipping the country
  9. A customer moving to a competitor
  10. A customer having an insurance claim

All these models are built in a similar way, but with different input data and outcome predictors.

Machine learning

What was presented above is a credit scorecard, we refer to these as traditional models. They are built utilising mathematical processes such as mathematical programming or logistical regression. These days an increasing number of lenders are opting for machine learning scorecards. These models are different in a few ways. Firstly, they are retrainable – that means models can be built and then rebuilt quickly. This is because we deploy them on a machine learning platform which prepares the data for us allowing quick retraining. Secondly the scorecards are a lot more complex utilising sometimes as much as 30 variables with multiple levels of segmentation. You can read more about our machine learning models – the advantage and disadvantages in our blog on the subject.

If you are looking for scoring services, please be in touch with us.

Principa offers the following:

  1. Building scorecards
  2. Validating/monitoring scorecards
  3. Re-aligning scorecards
  4. Setting scorecard cut-offs
  5. Deploying scorecards in our decision engine DecisionSmart
  6. Machine-learning models
  7. Scorecard mentorships

 

Thomas Maydon
Thomas Maydon
Thomas Maydon is the Head of Credit Solutions at Principa. With over 17 years of experience in the Southern African, West African and Middle Eastern retail credit markets, Tom has primarily been involved in consulting, analytics, credit bureau and predictive modelling services. He has experience in all aspects of the credit life cycle (in multiple industries) including intelligent prospecting, originations, strategy simulation, affordability analysis, behavioural modelling, pricing analysis, collections processes, and provisions (including Basel II) and profitability calculations.

Latest Posts

The Pros and Cons of a Multi-Bureau Strategy in Credit

Although not a new concept, very few lending organisations have deployed a true multi-bureau strategy (MBS). It is however talked about fairly regularly, but often dismissed as “too hard” or “not important enough”. So why should you consider a multi-bureau strategy? What are the key considerations? How do you go about deploying a MBS? This blog hopes to address all these questions.

PART 1: How to Cure the Post Pandemic “Collections” Symptoms

It has been a year and a half since the first case of the coronavirus (COVID-19) was reported from Wuhan, China. As we move into the third wave of the virus, there is an apparent dilution in both collection and recovery yields in the financial services sector, primarily because relief schemes and packages come to an end.

Predicting Customer Behaviour

Propensity modelling attempts to predict the likelihood that visitors, leads and customers will perform certain actions. It’s a statistical approach that accounts for all the independent and confounding variables that affects said behaviour. The propensity score, then, is the actual probability that the visitor, lead, or customer will perform a certain action.