What is the Probability of Default? Why is it essential for credit scoring?

As Chief Analytic Officer, I am responsible for the analytical efforts within all Creditsafe entities. Together with my great team of Data Scientists, we make it possible to develop scorecards which reflect the current credit risk conditions in the business world per country. We are responsible for the entire modelling structure of the scorecard, as well as the maintenance of the applied analytical methods and techniques.

By doing so, we contribute to customer-focused solutions where we can propose customer analytical solutions that support the customers in optimising their risk & credit management. Using the right data sets, a personalised analytical approach and the provision of accurate scores and PoDs, we help our customers minimise credit risks and enable them to make predictive decisions.

It is our common goal to ensure that clients can make correct business decisions based on objective and up-to-date data. This fact ensures that we can contribute to the strategic design of any organisation by putting the analytical aspect at the forefront. 

This is common practice for credit scorecards, if we want to ensure that credit scores are as accurate as possible, reflecting the current economic, social and political climate. Time marches on and so do the underlying elements such as data and the way legal company forms are defined. These are always subject to change.

As time passes, each model must be re-evaluated and adapted to current trends and behaviours.  This is why the PoD has become an important asset. The underlying statistics have changed, the scoring model has changed and, as a result, the meaning of the scores from 1 to 100 has changed.

From more than 8,000 data sources worldwide, we carefully identify the factors that have a strong correlation with company failures. To increase the predictability and stability of our credit score, we need to segment companies within the business population based on available information, size and type of business. Within each credit score card, the importance that each factor has on the credit score varies depending on its statistical relevance in relation to a business failure within a 12-month period. The main basis for the segmentation of the Belgian business population is the size and legal classification of the company.

The essential concept behind Creditsafe's scoring approach is to accurately predict the behaviour of companies (in terms of their good/bad performance over the next 12 months) using a set of characteristics, such as filed annual accounts and type of sectors, that clearly indicate why a company is considered high or low risk.

As already mentioned, the PoD is a mathematical calculation that indicates the probability that, in this case, a company will go bankrupt or become insolvent within 12 months. It is a simple and representative way for companies to accept, modify or even reject credit applications. This probability can be applied to a wide range of risk management and credit analysis scenarios. The PoD depends not only on the characteristics of the debtor, but also on its economic environment.

We calculate the PoD for each company separately based on their combination of features. To make the PoD comprehensible, we convert it into a simple score ranging from 1 (highest risk) to 100 (lowest risk).

The PoD is essential for calculating the individual risk of failure. This way, companies can get a better and more predictive insight into possible losses and risks in the near future, and take action based on this insight if necessary.   

The PoD helps risk and finance teams to tell how stable a company is and, more importantly, what the associated credit risk is to do business with. The Creditsafe scorecard provides a score between 1 and 100 for the highest and lowest risk respectively. This scale is created based on the outcome of the PoD, expressed as a percentage. The higher the Creditsafe credit score, the lower the PoD percentage and vice versa.

More and more companies today calculate their ECL or = Expected Credit Losses outcome within their IFRS9 framework (ed. IFRS 9 contains the necessary requirements for a new provisioning model that allows credit losses to be recognised earlier by companies). To do so, they need an individual PoD, as a change in risk itself is one of the stages within that framework.

For example, many large banking, insurance and leasing companies create and use their own scorecard based on individual PoDs, also in such a way that they can combine their own internal data (e.g. data on past due days) with information or data from external business data companies, such as Creditsafe. These types of companies and customers want to tailor Creditsafe PoDs to their own portfolios or data population. Moreover, they implement and use these PoDs based on their own definition and collection policy. It is easier to do this in a more granular way if they use the individual PoD instead of the traditional credit score, because the risk to the company is more accurate respectively.

An individual PoD also makes it possible to reconcile risks from different sources. Even if the company is a sole trader, it will be able to make a more accurate forecast of future credit losses, which also means that it can plan ahead more adequately.

Many of these companies are optimising and expanding their credit risk management as a result of the current economic climate. They are placing more and more emphasis on individual PoDs or an exact estimation of the risk, instead of an average risk per score.  In this way, they want to gain a better and more predictive insight into possible losses and risks in the near future.

Be precise, the desired result can be found in the details. Also, be periodic in the follow-up of your debtors, nurture yourself with the right insights and understand the changes in your portfolios. In addition, it is essential not to lose sight of the humane side. Therefore, regular contact and dialogue with your customers is crucial.

I can only recommend companies to support and empower organisations with anaytic insights. Don't just put your analysts in a corner somewhere in the office and expect them to deliver tables and diagrams. Instead, integrate them into the whole organisation by challenging your policies, identifying logical gaps and by collaborating with commercial teams.