By Ben Guy*
IFRS9 is the new accounting standard that replaces IAS 39. Up until now, it has required many organisations to dedicate significant resource in an effort to ensure readiness for the 1 Jan 2018 implementation deadline.
The standard includes a number of high level changes but the impact to impairment models is the most significant. In the UK and Europe this isn’t simply because of the significant increase in model complexity, but also because of the uncertainty created in the UK wake of the surprise Brexit vote outcome.
The issue with Brexit is no one knows what will happen. We don’t even know when the beginning of the process to leave will start. So how can anyone know, with any reasonable level of certainty, what the impact might be on something as fundamental to the UK economy as house prices or unemployment over the medium or long term? So far the economic forecasts don’t seem to have added much with the economy bearing up better than anticipated, but are the effects just delayed?
It’s this uncertainty that makes IFRS9 particularly challenging across a number of disciplines. No longer can we assume what happened in the past will happen in the future; we now need to incorporate macroeconomics such as housing prices or unemployment. At its purest, the standard requires us to weight the probability of different outcomes but many banks are poorly equipped to plan or justify their assumption on a probability-weighted basis.
Where are we right now Currently no institution has published any final numbers. Progress varies from one organisation to the next, but even those who are towards the end of their build phase are still going back to the guidance to improve their understanding and interpretation.
Before implementation, the models will need to be independently validated. Audit will play a significant role in both quantitative and qualitative validation and this will require a more technical level of expertise, bringing with it another challenge.
Risk modellers/developers are building and managing the models on an ongoing basis. On that basis they cannot validate those same models whilst maintaining any degree of independence. It’s this requirement that is (1) likely to require the engagement of third party suppliers to help provide a level of independent validation and insight otherwise unavailable internally, and (2) result in new skillsets and expertise in different roles across the industry, in particular audit.
What happens next? Qualitative validation will take place in advance of any model being implemented and this covers things like data validation, model methodology, documentation and an assessment of to what extent the current and future anticipated economic climate can be projected through the models. The ongoing quantitative validation covers calibration, model stability and discriminatory power. Calibration is updating the model to reflect the new information available, including the application of new macroeconomic scenarios. Model stability is about (1) looking at what we thought would happen against what did happen and (2) gaining an understanding of how the numbers move based on changes to the drivers, either within the organisation or the economy. Discriminatory power is about ensuring the right accounts or segments have the right level of provision to ensure appropriate business decisions are taken within each portfolio.
It would seem the most common plan within the industry is to begin parallel running of IFRS 9 models for comparison to IAS 39 as early in 2017 as possible. This will allow organisations to understand both the financial and operational impacts to their business. However, parallel run is unlikely to provide answers to the range of questions that the auditors are likely to be interested in, for example:
- How does the bank justify that Stage 2 criteria for identifying significant increase in risk, a reasonable and prudent view?
- How does the bank demonstrate that it is holding appropriate lifetime provision relative to peers on similar Stage 2 assets?
- How does the bank illustrate that it understands and has appropriate estimates of the key economic sensitivities in its portfolio on Probability of Default and Lifetime Expected Loss?
Today, there is an abundance of information with data available to everyone; but IFRS 9 is brand new. There is no peer-to-peer comparisons, no component benchmarking of models built on pool data, and no comparisons of peer modelled asset averages to benchmark models on the pool. The complex nature of IFRS 9 means comparing aggregate provisions or the individual component outputs, may be impossible without any of the above.
All of this highlights just some of the challenges ahead. Whether you are building risk models, managing risk functions, auditing or regulating them or investing in opportunities in the future; it’s a confusing a complex time for everyone but things will become clearer as we move forward and data becomes more readily available.
*Ben Guy is technical consultant, regulatory practice, 4most Europe Ltd, a specialist credit risk analytics consultancy.