InterviewIFRS9
- 1. interview training marcus evans. What do you think are the next steps on the evolutionary path of credit risk modeling? A. Roughly ten years ago, we had Basel II. It focused a lot on internal models and regulatory capital. There were incentives for developing IRB models however these were largely around stability of models and capital using the so called Through-the-Cycle (TTC) anchor. After the 2008 financial crisis, Stress Testing became an important tool which focused on tail losses and now IFRS9/CECL requires new, more forward looking ways for measuring losses. All this happened piecemeal, with Basel II becoming law in Europe and Stress Testing CCAR-DFAST becoming one in US banks. In next few years when teams address IFRS9 requirements in Europe and CECL requirements in USA, they will be left with pieces for all these capital requirements but no integrated view. Successful and cost-efficient capital management calls for an integrated framework around all of these connected requirements with end-to-end integration of data, models and systems. Institutions possess this information in form of PD, LGD, EAD and Stress Testing models as well as IFRS9 requirements. The next step is to connect all these pieces of the jigsaw puzzle together to get one overall view of capital. We see the need for such an integrated framework going forward. marcus evans. What is the single biggest modelling or methodology challenge that institutions face? A. Within the wholesale, corporate and commercial space, the model outputs, be it PD, LGD or EAD have all been designed over the past 10 years to predict mid-point. This led to model outputs which were through-the-cycle (TTC) like due to (a) capital stability focus in Basel II (b) lack of data and modelling techniques and (c) mimicking Agency ratings which are TTC-like. In fact, two research articles that we published in 2015 in Journal of Risk Model Validation show that Agency Ratings are around 80% TTC. So within this space, institutions which anchor their ratings to Agency ratings end up creating hybrid TTC-like internal models. This is a problem because IFRS9 calls for adjustment of current conditions and then a future forecast, therefore such TTC-like models would not comply with the new requirements. What institutions need is a Point-in-Time (PIT) / Through-the-Cycle (TTC) framework to convert one type of model output to the other. If institutions don’t do this they will face not only non-compliance risk but, more importantly the loss calculations could be off by a magnitude of 5 times. We see the need for PIT/TTC methodology as the single biggest challenge. marcus evans. What are some of the requirements that make IFRS9 challenging? A. The IFRS9 rulebook unlike the CRR rulebook is pretty small and leaves a lot to interpretation. Take this one for instance – “an unbiased and probability- weighted amount that is determined by evaluating a range of possible outcomes”. This innocent sounding one line is hotly debated. For mathematicians like us, some things are obvious but those not trained in thinking in terms of probability distributions, it is a challenge. We had a client saying that an economist will select scenarios but the economist himself confessed that his selection will be biased with his personal views and downward biased after years of seeing baseline and stress scenarios as part of Stress Testing. Designing an upward scenario is not what economists have done lately so how can the selection of few scenarios selected by one person be unbiased? The economist later confessed that he would not be comfortable with assigning probability distributions to scenarios. We can talk about this very topic for hours but this just shows you how requirements are hotly debated. marcus evans. With all this debate around addressing modelling challenges, are institutions investing enough time in implementation? A. We definitely agree that institutions are not leaving enough time for implementation. In some of the methodology discussions, institutions do not even consider implementation issues, which surprises us. In our opinion firms need to adopt batch processing so that IFRS9 loss calculations can be done efficiently and accurately with minimal human intervention. As a final output we envisage an analytical engine with makes use of institution’s raw data and existing Basel II and Stress Testing model output, converts it all to IFRS9 output with automated calculations and can run several scenarios and loss calculations in minutes, month-after-month which would mean minimal operational costs. For this reason, we always present an end-to-end view of modelling, methodology and implementation as we are doing in this course. marcus evans. How will the course you designed help address these challenges? A. The two-day IFRS 9 course that we will be delivering in multiple locations provides an end-to- end view of challenges and solutions. It summarizes all the debate and discussions till date and is also a great way for delegates to benefit from our decades of global experience. However, with the clock ticking, it is time for action. Delegates should start with small projects and start designing prototype solutions with the help of advisors. No solution will be perfect but with a couple of iterations, a tactical working solution can be achieved in a matter of few quarters and a full blown solution in a year’s time, in time for global rollout. About the event This two-day course will review the IFRS9 provisioning rules and describe ways of producing unbiased probability weighted estimates of ECLs, building upon an institution’s existing models. The workshop will provide insights into challenges of developing compliant solutions. Techniques for satisfying novel requirements like “significant deterioration” and “unbiased probability weighted forecasts” will be discussed. At the end of the course, delegates would have seen and discussed various tools and techniques which will help them design IFRS9 solutions that work for their own institutions. Footnotes: The presented opinions and methods are solely the responsibility of the course expert trainers and should not be interpreted as reflecting those of Deloitte. The credit risk modeling industry has faced numerous regulatory challenges over the post financial crisis years, and IFRS 9 is yet another complex and far reaching piece of regulation which will have a dramatic impact on the way in which banks model their losses in this area. With implementation of the wider regulation now firmly in sight for 2018, banks are determined that this year will see them tackle the modelling challenges so that the remaining time until implementation can be devoted to testing and practical issues. However, the modelling questions are by no means straightforward and with a lack of clarity in many areas, the industry is looking to benchmark approaches to allow them to get on with the practical challenges. Ahead of the IFRS 9 Models, Methodologies and Implementation 2-day course, we spoke with Scott Aguais, Founder and MD and Gaurav Chawla, Senior Consultant of Aguais and Associates (AAA, an affiliate of Deloitte) about solutions to modelling, methodology and implementation challenges faced in IFRS 9 modeling. They emphasize the need of an integrated framework for designing IFRS 9 solutions and recommend tailored solutions for each institution’s existing data and models. An Integrated Framework for addressing Credit Risk Modelling Challenges posed by IFRS 9 Intervew with: Dr. Scott D. Aguais and Gaurav Chawla Gaurav ChawlaDr. Scott D. Aguais For more information about the event, please contact: Gareth Banks, Tel: +44(0)20 3002 3400 Email: GarethB@marcusevansuk.com www.me-financialtraining.com