Request a Brochure

Learn, explore and unleash your inner chef.

For more information, contact:
Yiota Andreou
yiotaa@marcusevanscy.com

How banks are moving from their existing Basel Models to meet the IFRS9 demands?

During the initial phase of working through the final standards last year, there were still serious considerations of developing different model frameworks for IFRS 9 compared to Basel A-IRB models. Now this question seems to be settled: as long as you have Basel models at hand, satellite-type of models are increasingly used for IFRS 9. This approach makes sure that both worlds remain reconciled and the effort to develop and validate models does not increase excessively.

What are the differences and expectations between IRB and IFRS 9 modelling?

On the technical side, there are different requirements to be fulfilled for IFRS 9 impairment compared to Basel. Where Basel focuses on long term averages, IFRS 9 is targeting forward-looking point-in-time models, considering most recent economic forecasts. IFRS 9 requires lifetime estimates, they might be prudent but not outright conservative, and the use of eligibility criteria or cutoffs are basically a “no go” for IFRS 9. On the structural side, regulators and standard setters have very different objectives: Regulators focus on the high quantile (99%, 99.9%) of the loss distribution – the rare event of a crisis is driving capital requirements and the RWA formula. Standard setters focus on the average outcome, based on “probability weighted outcomes of potential future scenarios”.

Could you please briefly explain how to measure and calculate expected credit loss?

It is common practice to describe ECL (expected credit loss) by its components PD, EAD and LGD. For 12 month outcome period, ECL is the simple product of these terms (without consideration of potential dependency structures). For lifetime ECL, the term structure of all three components has to be estimated and aggregated. The real challenge is the quality assurance of the aggregated lifetime ECL estimates – for proper back-testing sufficiently long time series are required.

What are the differences of using PD and LGD under Basel and IFRS 9?

As already mentioned before, Basel focuses on long term averages, whereas IFRS 9 needs forward-looking point-in-time estimates. For PD, lifetime concepts are needed on top of the standard 12-month outcome period generally used for pillar 1 or 2 models. For LGD, the discounting regime has to be changed – instead of the “economic-loss”-based approach in Basel (considering impact of collections cost and the time value of money through application of an adequate discount rate), the original effective interest rate has to be used. Additionally the LGD needs to be based on any future loss event, not just within next 12 months – which means that anticipated changes in market values etc have to be considered appropriately.

What would you like to achieve by attending the 4th Edition Credit Risk Modelling under IFRS 9 Conference?

I’d like to get an updated understanding of the current status of the industry in the development and implementation of IFRS 9 methods, especially on loss estimation, bucket transition and economic forecasting. And to hear about recent comments and interpretations by regulators or standard setters, discuss how they are going to be handled by banks and which type of solutions are being considered. And maybe we get some practical presentations how banks handled the start of the parallel run.

Ahead of the 4th Edition Credit Risk Modelling under IFRS9, read here an interview with Mr. Wolfgang Reitgruber, FVP, Group Credit Risk Modelling at UniCredit s.p.a.

About the conference:

This marcus evans event will guide financial institutions through revealing key market knowledge of the solutions to the challenges posed by IFRS9 on credit risk models and enable institutions to share positive and dynamic approaches in time for the 2017 practice year. The 4th Edition Credit Risk Modelling under IFRS9 Conference will take place from the 23rd until the 25th of November 2016 in Vienna, Austria.

About the speaker:

Wolfgang is responsible for model development governance of credit risk methods (rating, scoring, EAD, LGD, risk mitigation) in UniCredit s.p.a. He also works as independent researcher since November 2012, in the area of credit risk related methods on basis of regulatory (Basel II/III A-IRB) and accounting (IFRS 9 impairment) standards. Primary achievement was the introduction of the EL backtest, as published in the Journal of Risk Model Validation¹ . He currently works on further extending the Impact of Risk© framework to support development and validation of life-time expected loss models in an IFRS 9 environment ².

(1) “Expected loss and Impact of Risk: backtesting parameter-based expected loss in a Basel II framework”: Journal of Risk Model Validation, Vol. 7/3, pp 59-84, 2013.

(2) “Expected loss provisioning under upcoming IFRS 9 Impairment Standards: A new source of P&L volatility — can we tame it?”:forthcoming in: Journal of Risk Management in Financial Institutions, Vol. 9/4,pp 332–343, 2016.

How banks are moving from their existing Basel Models to meet IFRS9 demands?

An interview with Wolfgang Reitgruber

**The presented opinions and methods are solely the responsibility of the author and should not be interpreted as reflecting those of UniCredit s.p.a. or UniCredit Bank Austria AG. All rights reserved.

Practical Insights From:

Wolfgang Reitgruber

Vice President‑Group Credit Risk Modelling

UniCredit

Diana Kapsa

Head of Credit Methodology‑Retail

UBS AG

Rik Van De Weerthof

Head of Credit Risk

Nationale‑Nederlanden

Louise Lindgren

Chief Risk Officer

Länsförsäkringar Bank

Hannes Kazianka

Senior Examiner

Oesterreichische Nationalbank

Michael Perrie

Head of Model Validation

Investec

Wolfgang Reitgruber, FVP, Group Credit Risk Modelling, UniCredit s.p.a.