The Data Analytics Blog

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

All Posts

Economic Modelling Guidelines In IFRS 9

July 6, 2018 at 9:01 AM

As part of the group that was the second company worldwide to become IFRS9 compliant, IFRS9 has been at the forefront of what we do.  We have assisted nearly 20 companies on their IFRS9 journey over the last two years.  This blog forms part of a more extensive series on IFRS9.

In a previous blog, we highlighted some tips for managing manual overrides for IFRS9 models.  In this blog, we explore some of our learnings in economic modelling with IFRS9.

Part of IFRS9 is the creation of data-driven models to most accurately predict expected credit losses.  Much of the prediction is derived on historical data, but IFRS9 also requests that we need to consider forward-looking projections or considerations, as the future may not be exactly like the past that is often the underlying premise of forecasting models.  Hence the requirement for expert adjustments with management foresight.  These expert based adjustments could be in the form of:

  1. economic outlook forecasting with potentially an analytical model that defines the complicated relationship between economic indicators and portfolio performance or
  2. management judgement of policy, strategy or environmental events that would not be reflected in historical data and thus not in the expected credit loss outlooks

IFRS 9

Economic forecasting and credit loss relationship modelling

Below we summarise some of the lessons learnt by Principa within the three primary considerations in building a forward-looking economic indicator adjustment modelling solution:

Macro-Economic Indicator Forecasting and Outlook Scenarios

macro

While modelling might happen on historical actuals, the model execution would be on economic forecasting to create economic scenario's.  Modelling on the historical forecast is an option, but then you run the risk of forecasting sentiment changes impacting the model.

Source a credible forecast, potentially from an independent 3rd party.  Internally sourced, requires evidence of approach and outcome to evidence objectivity and rational.

At minimum define three possible scenarios (with possibly a strictly data-driven base scenario) with the associated likelihood of each situation.

Modelling the Relationship between Macro Economic Indicators and the Components of Credit Losses

Modelling

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Models would look at the key credit performance indicators that drive credit losses such as modelling the relationship for the probability of default or loss given default separately.  One can, however, go even deeper to identify key leading indicators such as flow into collections or debt sale price. Consider that both timing and magnitude could be significant.

It is important to note that the outcome variables are net of mitigating actions. So often the relationships might prove weak or counter-intuitive.   As you are looking to create scenarios with logical ranking, avoiding illogical relationships within your model is advised.

Also, consider lagged relationships.  E.g. it might take 6 to 12 months for suppressed economic growth to wash through to unemployment or other constraints on disposable income and then for the effects to wash through of credit performance from early stresses to actual losses.

Lastly be careful to double count the macroeconomic impacts.  In other words, consider macroeconomic effects already embedded in outlook.  At Principa we look to avoid this by modelling the relationship of macroeconomic indicators to the model error with the underlying assumption that the model would be perfectly accurate if the environment did not change.

Economic indicators could include:

  • FX rate,
  • Interest rates, 
  • Economic growth,
  • Inflation, 
  • Petrol inflation,
  • Food inflation,
  • Indebtedness,
  • Disposable income,
  • Saving ratios,
  • Unemployment,
  • Confidence indices

While these could be expansive after also considering various lags and permutations, consider reducing the variables through considering the correlation between all economic variables, removing the variables where correlation is high, and removing variables with insignificant coefficients.

Ensure that the models follow a robust development methodology with detailed modelling write-ups and model validations and back-testing.

ECL Procedure to Provide the Scenario Weighted Outcomes

ECL

IFRS9 requires that forward-looking overrides are scenario weighted and as such requires that an organisation to:

  • Create the capability to run expected credit loss overrides in line with the modelled scenarios and related performance influences
  • Create the capability to output a single scenario likelihood-weighted outcome of expected credit losses.   It could, however, be a single scenario weighted override by model component incorporated in the final expected credit loss calculation procedure.

Forward-looking overrides through management judgement

Forward-looking management judgement should be exceptional and reserved for “out of the ordinary” events that would not be reflected in historical data and thus not in the expected credit loss outlook even after considering the scenario weighted economic indicator overlays.  We covered key considerations of management judgment in a previous blog.

Principa offers a wide-range of products and services relating to IFRS9.  These include analytical, advisory, model templates, reporting packs, deployment support, model software, etc.  We’ve also been able to help companies post-IFRS9 to make the most of their modelsContact us to find out how Principa might be able to help your business.

Get IFRS 9 Compliant

Edwin Cross
Edwin Cross
Prior to joining Principa in December 2013, Edwin had 13 years’ consumer credit (credit cards, revolving credit and personal loans) experience in Barclays Bank and Standard Bank SA with 10 years’ specialised in credit risk management across South Africa and Europe in particular United Kingdom, Spain, Germany and Italy of which over 3 years with Chief Risk Officer accountabilities. Edwin is considered an expert across the full customer life cycle and has lead and implemented a number of highly successful credit, marketing, collections, scoring and provisioning initiatives. Since joining Principa, Edwin continued to successfully deliver risk-based and profit-based initiatives to key clients (specialised lenders, large retail banks and retailers). He holds a M.Comm degree from Stellenbosch University.

Latest Posts

[Slideshare] How To Make Your Business Data Work For You

Common barriers to success: Skills shortage: data scientists are in high demand and in low supply. Companies lack the skills to develop advanced data analytics or machine learning applications. Cost: recruiting and building up or training a team, as well as infrastructure costs are immense. Inefficiency and low ROI on: acquisition campaigns; re-activation and retention campaigns; outbound sales calls and debt collection. Resulting in: No or ineffective use of data. High cost to get insights from data. Low returns from campaigns. What’s the alternative? Machine Learning as a Service (MLaaS): removes infrastructure skills and requirements for machine learning, allowing you to begin benefiting from machine learning quickly with little investment. Subscription based pricing, allowing you to benefit using machine learning while minimising your set-up costs and seeing returns sooner. Answers as a Service: Use historic data and machine learning to allow answers to increase in accuracy with time. MLaaS with predictive models pre-developed to answers specific questions: Genius Call Connect: What is the best time and number to call customers? Genius Customer Growth: Which customers are most likely to respond to cross-sell? Genius Re-activation: Which dormant customers are worth re-activating? Genius Customer Retention: Which customers are most likely to churn? Genius Leads: Which contacts are likely to respond to my campaign? Genius Risk Classifier: Which debtors are most likely to pay or roll? Benefits of Genius: Quick and cost-effective ability to leverage machine learning: Minimal set-up time Minimal involvement from IT Subscription based service Looking to make your data work for your business? Read more on Genius to see how it can help your business succeed. 

5 Must-Join Facebook Pages For Data Science, Machine Learning And Artificial Intelligence In 2019

While LinkedIn has traditionally been thought of as the business or work focussed social platform, Facebook has been making headway into gaining market share in the space as well. With company pages and groups, Facebook is catering to every interest and aspiration that people might have – and combining that with their social interactions and news sources. Facebook aims to give users a one-stop-shop experience, and it’s very good at doing it.

Our 2018 Customer Acquisition And Engagement Blog Roundup

Our final roundup this year covers two of our main topics: customer acquisition and customer engagement. We’ve not covered these topics in depth this year, and so decided to combine these two to provide a roundup of the best of both.