By Silvio Santarossa, Partner - Risk Advisory Services
and Kristian Lajkep, Regulatory Compliance Officer
The European Banking Authority (EBA) has conducted a second impact assessment of IFRS 9. The project was carried out between November 2016 and February 2017 and the results were disclosed to the public in July 2017. Similarly to the first exercise, the project was performed on close to 50 institutions across the EEA and was focused on assessing the preparedness of financial institutions for the new standard, the impact of regulation on regulatory own funds and other prudential requirements. The general state of preparedness of institutions has improved considerably, both in terms of models and processes, but also in terms of quality of information. However, the report as well as our own experience in IFRS 9 related projects (mainly linked to the impairment model stream) across Europe, give rise to some interesting observations and comments.
Banks were asked to state their degree of preparedness, ranging from early design, advanced design, to building and testing. Most banks identified themselves as being in the building stage with regards to classification, measurement and impairment.
However, only few banks have reached the testing phase. As this sample of 50 banks is too small to be a reliable inference, it would be prudent to assume that a substantial number of financial institutions are lagging behind and still need some complex items of IFRS 9 to be addressed.
Major consequence: Whilst institutions will struggle, they will probably manage to be ready in time with regards to implementation. We can reflect on the possibility of the buffer, initially intended for testing and/or parallel run, being used to catch up the delay instead. This may increase the risk of non- propriate testing or inadequate anticipation of final financial impacts.
Smaller banks tend to be less advanced in their preparations, though they have made considerable progress. For the most part, even the small banks have provided all necessary information on the estimated impact of IFRS 9.
Whilst most banks are on the right track, the EBA is concerned that some banks, particularly the smaller ones, are progressing too slowly to be ready for the implementation of IFRS 9.
A majority of banks have stated that they are not planning to apply IFRS 9 hedge accounting requirements, but would rather stick to the current IAS39 hedge accounting requirements. Those that are, do not regard this as a priority and are in early design state only.
In some cases, the stakeholders of the companies do not get sufficiently involved in the preparation for IFRS 9. This may have adverse consequences both regarding the allocation of resources to the task and their own ability to exercise their duties during and after the transition to IFRS 9.
A famous movie actor would comment this as follows: “What else?”. Indeed, the above list concerns little else than the main changes brought by IFRS 9 compared to IAS39 (if we consider that some aspects refer to lifetime modelling as well).
Nearly everyone struggles with at least some of the IFRS 9 challenges. Some of these challenges are significant enough to necessitate the use of various shortcuts, simplifications and assumptions. Even if the EBA advocates the use of strong governance and methodologies to monitor those implifications, we can expect a large variance of methods across banks, which will make the comparison quite tricky as of 01/01/2018.
Taking as an example the assessment of a significant increase in credit risk, as also stated in the report, banks will use quantitative and/or qualitative elements to assess it. However, there is no reason to believe that all banks will use the same criteria for exactly the same product or category of clients (or the same level of thresholds for the quantitative part).
For the bank itself however, it would be highly recommended to test the extent of sensitivity of ECL measurement to each scenario or each decision tree defined as the set of triggers for a significant increase in credit risk. A too high volatility might question the assumptions taken so far.
Generally, the implementation of IFRS 9, on the one hand, leads to a decrease in Common Equity Tier 1 ratio (by 45 bps) and has a similar effect on the total capital ratio (a decrease by 35 bps). Both these decreases are not as bad as in the previous exercise, but remain substantial. The volatility of profit and loss, on the other hand, is going to increase - at least that’s what 72% of the banks believe. This is mostly due to moving exposures from stage 1 to stage 2 and including forward looking information, and the resulting need to reassess each reporting period – leading to a variance, in figures.
Using the standardized approach (SA), smaller banks have calculated a larger impact on their own fund ratios and a smaller increase in provisions (13% on average) than their larger peers, which tend to use internal ratings-based approach (IRB). The total estimated impact on the own fund ratios is mainly due to the impairment requirements and, to a somewhat smaller degree, to classification and measurement requirements. The main driving force behind this impact is the estimation of lifetime expected credit losses (LECL) for stage 2 exposures. Compared to banks using IRB, banks in SA cannot benefit from the capital adjustment linked to the excess/shortfall of provisions (accounting ECL vs regulatory ECL).
In the first exercise, the provisions increased more than in this one. The reduced rate of this increase is most likely due to a greater degree of preparedness of the institutions as well as possibly better macroeconomic conditions.
There is a visible reduction in and occasional absence of planned parallel runs of IFRS 9 and IAS39 compared to the last period. The EBA has expressed concern over this phenomenon and advises those banks that plan for no parallel runs to at least rigorously test their IFRS 9 processes and approaches, particularly with regards to impairment.
The EBA also hopes that the application of IFRS 9 will lead to both institutions and regulators having access to better information as to the sort of data, assumptions and simplifications that different banks employ for the calculation of expected credit losses (ECL), enabling them to make a more disciplined use of these. So far, the quality and availability of historical data and the assessment of ‘a significant increase in credit risk’ seem to be the greatest thorns in the back of most institutions, so some simplifications would be needed. The EBA stresses the importance of a sound judgement when making them. The consistency of the key assumptions across all banks seems to be also important.
The EBA stresses that it is absolutely crucial for all relevant key stakeholders to ensure a timely ongoing dialogue in relation to the implementation of IFRS 9. The EBA will do its utmost to maintain this dialogue, as well as to encourage any effort that may result in an easier implementation of IFRS 9 or a better quality thereof.
In the longer term, the EBA would like to enhance its and other institutions’ knowledge of different implementation practices. For that end, it proposes a review of the implementation of the EBA Guidelines on ECL as well as a study on the impact that different inputs, models and methodologies would have on ECL. A discussion at the European and international level (with the Basel committee for instance) should also be maintained in order to see whether the interaction of the capital framework with the ECL model for accounting can be improved with changes to the regulatory framework.
As mentioned earlier, banks will use – to a certain extent - some simplifications to overcome some IFRS 9 hurdles or data issues in order to reach the deadline of 01/01/2018. Banks will then probably continue improving key points of the implementation of IFRS 9 after its initial application.
In addition, for banks under SA, the time and energy already invested to cope with the IFRS 9 challenges (a.o. with PiT models) could represent a solid foundation for a transition to the future IRB application.
A significant number of banks have yet to decide on their processes for the validation and back-testing for the ECL measurement. Others intend to rely on the existing validation processes for prudential models. The latter might not be sufficient in the context of IFRS 9, so the banks will probably need to develop more robust validation processes and a monitoring system capable of capturing the complexity and key elements of the ECL models, to ensure a high-quality application of the standard.
Finally, at this stage, there is no obvious evidence that IFRS 9 will dramatically impact the lending practices and business models of financial institutions. Only the future will provide more insights about it.
Nevertheless, it is already certain that the regulator will put a strong focus on the post implementation review. In the short term, this will be visible already next year as the 2018 EU-wide stress test exercise will take into account the implementation of IFRS 9.