Friday, 5 February 2016

Accounting for expected credit losses under IFRS 9: Thoughts on benchmarking, calibration and the economic value

After the release of the new impairment standard in July last year, financial institutions started to crunch numbers to get a feeling of the impact of expected credit losses on their allowances. On the one hand, inclusion of lifetime losses in accounting for financial instruments is a logical and consistent step towards adequate consideration of changes in riskiness. It is already now the basis for evaluation of impaired assets. On the other hand, the net value per IFRS 9 accounting standards at day one is lower than the economic value. For (performing) loans which are sold at fair value, an immediate loss at day one has to be reported.

For better understanding of the accounting value in comparison to an economic benchmark, the final standard needs to be seen in the light of its true intent. The IASB reconfirmed in the “Basis for Conclusions”, that IFRS 9 impairment was the result of an iterative process to approximate an ideal accounting model for expected losses under operational constraints:

BC5.88ff: In the IASB’s view, the model in the 2009 Impairment Exposure Draft most faithfully represents expected credit losses … but [respondents] said that the proposals would present significant operational challenges … These operational challenges arose because entities typically operate separate accounting and credit risk management systems. BCIN.13 In response, the IASB decided to modify the impairment model proposed in the 2009 Impairment Exposure Draft to address those operational difficulties while replicating the outcomes of that model that it proposed in that Exposure Draft as closely as possible.

With this intent in mind, how can we derive a benchmark for the accounting value under the upcoming impairment standard? A valid benchmark to reflect principles of IFRS 9 has to have the following properties: 1. It shall fully reflect changes in the riskiness of the borrower, as measured by changes to the expected credit loss. 2. It shall be neutral with respect to other (mostly external) impacts, such as changes to capital requirements, profit expectations, yield curves, interest rates or market sentiment. 3. The effective interest rate method shall be applied to amortize cost and income over the life of the loan.

The discounted cash flow (DCF) method with a fixed discount rate (reflecting the required neutrality with respect to external impacts) gives such a benchmark, when using expected cash flows, corresponding to contractual ones net of expected losses, as basis for the evaluation. Due to its underlying principles we can call this benchmark “idealized Amortized Cost Value”, or iACV©. This value actually corresponds to the net carrying amount as outlined in the Exposure Draft 2009.

The figure shows the typical situation for a non-amortizing loan, term 5 years, and constant gross carrying amount of 100. The y-axis denotes the net value after allowances, whereas time in years is depicted on the x-axis. Expected credit losses increase gradually with time, causing transition to bucket two after 1.5 years in this example. The smooth curve describes the benchmark iACV© as defined above. The other curve describes the net accounting value according to IFRS 9. It starts more conservative (caused by the 12-months ECL allowance), but gets non-conservative with gradually increasing risk level. Transition to bucket two causes the accounting value to become conservative again, at a much more significant level than in bucket one.

The development of iACV© allows for a continuous quantification of the impact of changed lifetime loss expectations on a credit portfolio, without artificial cliff effects. Any changes to allowances on top of this benchmark, quantified by the difference between the accounting value and iACV©, are caused by specific choices when implementing the local IFRS 9 architecture (primarily the transition process between buckets one and two) and the approximative nature of the final standard.

When evaluating the local implementation of IFRS 9 impairment, this deviation should be among the key indicators: It reflects the closeness to the primary intent of IASB. High positive deviations indicate potentially hidden reserves. Large negative deviations may indicate hidden liabilities. If you want to be ready for the upcoming challenges by auditors and regulators, you should ask yourself the following key questions:
  1. Do you have the necessary steering tools in place when drafting and fine-tuning the transition process from bucket one to two?
  2. Are you prepared for discussions with auditors and regulators to justify the level of allowances derived on basis of your local IFRS 9 architecture?
  3. Can your risk systems monitor iACV© to ensure full visibility of ECL impacts on an ongoing basis?
If you are interested into these and other considerations around IFRS 9 impairment, please follow me on LinkedIn, send me an email or join the 2nd Annual Credit Risk Management Forum in Amsterdam 19th and 20th May.

The presented opinions and methods in this presentation are solely the responsibility of the author and should not be interpreted as reflecting those of UniCredit Bank Austria AG and CFP.

Written by:
Wolfgang Reitgruber
Head of Credit Risk Modelling, UniCredit S.p.A., Austria
Senior Advisor, Walkergasse 25, 1210 Vienna, Austria

No comments:

Post a Comment