Could you please expand on how to use multiple economic scenarios to calculate expected credit loss?

For risk management purposes, every bank uses multiple economic scenarios. It forms part of the risk management armoury. However, the risk management purpose is to identify risks and risk drivers not captured by other more conventional tools like VAR. This is different from what the IFRS9 demands. For example, there is no need to use best or worst case scenarios and one must assign probabilities for each scenario. Instead, the standard stipulates an unbiased and probability weighted calculation. Implementing this requires many steps and strong governance. One has to identify the macro-factors that are affecting his credit portfolio and then define the scenarios and the time horizon of each. Next probability weights would have to be calculated or guess estimated. The impact on staging is a further complication that is not present in normal scenarios. The technical complication is very high and for banks that have limited resources or idiosyncratic business plans, the burden would be significant.

What are the conceptual problems with IFRS 9 and what can banks expect from one-off capital hits in 2018?

When Basel issued their guidelines for capital calculation they took into account the fact that smaller lenders may not have the in house capabilities for advanced modelling techniques. By contrast there is no such provision for IFRS9 ECL calculations. In addition, even though the standard requires a multiple scenario approach it does not get into any detail or approach. It seems that judgment would be the word most widely used when doing these calculations. This means that the capital hits banks would suffer would be uneven and non-representative of their underlying risks. For some portfolios that exhibit non-linearity, doing this calculation would be very costly. This may result in a redefinition of many business lines. This would create a systemic problem as more market participants align their business models and fewer are taking non-linear positions.

How can a bank maintain consistency in macroeconomic forecasts?

There is a variety of sources for scenarios and macroeconomic forecasts. One could use for example the scenarios/forecasts proposed by the regulator but these are based on the assumption that the bank is exposed mainly to macroeconomic events particular to that country. In other words, although they may be consistent for a country forecast, they are not appropriate for banks that are globally exposed. A way out of this is to use the assumptions that are built in the ICAAP submission. Getting the probabilities of each scenario consistent involves judgement, especially for forward looking estimates in the middle of Brexit. The Brexit scenario is a typical case of a “known unknown”. We are pretty sure that is going to happen but have no clue on how the systemic change would affect the economy and the banking sector. Nor do we know what exit deal would be struck but we need to forecast the effects of that unknown deal.

What are the best practices on validation and backtesting models and who should be responsible for backtesting these models?

Backtesting models is a very nice and easy to understand theoretical concept. However, as always, the devil is in the details. The IT, data, and modelling issues required for even the simplest backtesting are considerable. Backtesting should ideally be done by a separate unit from the one that is implementing it. As this is not always possible due to lack of knowledge or skills or other resources, the best safeguard is to have a robust governance process in place and accountability at every stage of the decision process.

What would you like to achieve by attending the 5th Annual Credit Risk Modelling under IFRS 9 Conference?

Sharing information and trying to identify best practices and intelligent practical solutions for the many conceptual and practical problems posed by the implementation of the new standard would
be, I hope, a major deliverable of the conference. The industry needs more information sharing and transparency if we are to serve our shareholders and fulfil our obligation to society.