LSIs: keeping up with IFRS 9

The global financial crisis exposed weaknesses in the accounting standards for financial instruments. International Financial Reporting Standard 9 (IFRS 9), which came into force on 1 January 2018, has introduced new approaches to impairment and to the classification and measurement of financial instruments. The new impairment approach seeks to address the issue of loss provisioning being “too little, too late” under the incurred loss model by replacing it with an expected credit loss model. According to this model, banks must build up provisions well before risks turn into actual losses so that they are more resilient and better prepared for crises and economic downturns.

The consistent application of all relevant aspects of IFRS 9 is key to ensuring that expected credit losses are calculated in an accurate and timely manner and that financial instruments are classified and measured properly. However, banks reporting under IFRS have faced and continue to face a number of difficulties when dealing with the new standard. Given the interaction between IFRS 9 and banks’ regulatory capital, ECB Banking Supervision and the national competent authorities have taken a keen interest in the implementation and application of IFRS 9 by significant and less significant institutions (LSIs).

In 2017 supervisors from the ECB and the national competent authorities conducted a thematic review on IFRS 9, assessing the extent to which significant institutions and LSIs were prepared for its introduction. Around 25% of LSIs report under IFRS; the others apply their national accounting standards. For LSIs reporting under IFRS, the review found that a number of banks had room for improvement in their preparations. A year on, supervisors examined the progress of 58 LSIs in implementing the new standard, taking into consideration the guidance issued by the Basel Committee on Banking Supervision and the European Banking Authority (EBA).

This latest review found that project governance of IFRS 9 application had improved in the course of 2018, as a higher number of LSIs involved all relevant management board members in the decision-making processes and provided better quality documentation to supervisors. Furthermore, most banks had more rigorous governance and internal control processes in place and third-party products were better integrated into the IT landscape.

Under IFRS 9, the classification and measurement of financial assets are determined by the outcome of two tests, the first of which assesses the objective of the business model used. Supervisors noted that an increasing number of LSIs involved governance and risk management functions in decisions about business model classification. However, internal policies needed to be clearer with regard to the reclassification of financial assets owing to changes in business models. Moreover, not all banks explicitly and consistently defined the acceptable significance and frequency of sales of financial instruments classified under specific business models.

The second test assesses whether the contractual cash flows of financial instruments can be considered solely payments of principal and interest. The majority of LSIs in the sample met the supervisory expectation that the solely payments of principal and interest test be conducted on all types of financial instruments in a standardised manner. An area for improvement was to have clearly defined benchmark tests to see whether instruments with a so-called modified time value of money met this test.

The IFRS 9 impairment model follows a three-stage approach to reflect changes in credit quality since origination or purchase. The definition of default determines the boundary between exposures with a significant increase in credit risk (stage 2) and credit-impaired exposures (stage 3). The banks in the sample had made progress in aligning accounting and regulatory definitions of default. In addition, most of the banks applied materiality thresholds for assessing defaults. Nonetheless, conditions for transferring exposures out of default (and thus out of stage 3) could be defined more precisely.

On the whole, the concept of “significant increase in credit risk” was sufficiently well defined when it came to qualitative and backstop indicators. However, practices for establishing the thresholds for stage transfer varied among banks. While the principle-based nature of IFRS 9 allows for a certain degree of flexibility, banks’ practices needed to be assessed to understand whether the variations were justified by differences in the banks’ risk profiles and the information available to them. Additional points of supervisory attention were the use of watch lists as a backstop indicator and clear rules on cure periods, which allow exposures to be transferred from a lifetime expected credit loss measurement (stages 2 and 3) to a 12-month measurement (stage 1). Supervisors also noted that not all banks had sufficient documentation to justify their use of the low credit risk exemption.

The incorporation of forward-looking information is a vital feature of the expected credit loss model. Most banks in the sample were able to assess the existence of a linear or non-linear dependency between the distribution of expected credit losses and macroeconomic scenarios, and base their usage of probability-weighted scenarios on this assessment. Moreover, significantly fewer banks used a time horizon of more than three years in the scenario analysis because predictions beyond this horizon would be difficult to make.

Banks usually conducted model validation and back-testing after expected credit loss models had become stable. Hence, validation frameworks had improved over the course of the past year. However, some fine-tuning of certain governance aspects was still necessary, in particular to demonstrate the independence of the internal validation function from the model development function. Moreover, many banks still relied on external expertise until internal validation was fully set up.

Finally, the lifetime expected credit loss calculation was analysed. While most banks had adequately and proportionately transposed the EBA guidelines on expected credit loss into internal policies, not all banks applied forward-looking information to the estimation of loss given default throughout the remaining lifetime. Furthermore, some banks estimating an expected credit loss very close or equal to zero lacked clear and consistent documentation to justify this.

Overall, the latest review found that LSIs had made substantial progress in implementing IFRS 9 and successfully incorporating the standard into their day-to-day business practices. However, there were still a few points requiring attention.