Now that IFRS9 is live and banks have turned their attention to the on-going validation and monitoring of their underlying models, what are some practical issues arising as a result of the staging calculation? Here, Antonie Jagga of PwC Singapore explores.
Significant increase in risk? – maybe not
At any one point, the bank would have applied its staging criteria, flagging those accounts it believes have experienced a significant increase in credit risk (SICR) since origination. A large proportion of these accounts would be flagged as SICR despite having experienced no actual change in delinquency status.
So what happens when a couple months down the line, there has actually been no deterioration in risk? At a practical level this means that the bank’s staging criteria are not working as intended, or the underlying scorecards need recalibration. Too many of these errors will mean that a bank is likely to experience volatility in provisions, holding either too little or too much provisions at any point in time. As part of the validation of SICR criteria, banks should regularly back-test the appropriateness of its staging criteria, and establish an acceptable threshold for these classification errors.
Wait, the economy didn’t turn out as expected?
In theory, the inclusion of forward-looking elements in a bank’s staging criteria should reduce volatility as expectations of future changes in risk should already be fully included in lifetime probability of defaults (PDs). For example, where a negative macroeconomic outlook is incorporated in the staging, a higher proportion of Stage 2 accounts should be expected. As time passes, and these conditions materialise, more delinquent customers should be observed, in line with the staging expectations.
The main challenge with the inclusion of macroeconomic factors in staging is that the forecasting of these in its very nature is not very accurate, especially as the forecast time horizon is extended into the future. If banks were accurately able to predict the future state of the economy, we would see significantly more (and more profitable) market activities.
From a practical perspective, there has been a move to simpler forward-looking models, and less material statistical linkages between economic conditions and model parameters for SICR classification. The governance process around the setting of macroeconomic scenarios, ensuring consistency and repeatability, is also increasingly important.
What do you mean this customer is high risk – we need to sell more to this group!
Finally, there is a natural tension between business and risk when flagging SICR customers. It is important to ensure that the SICR criteria are aligned to existing credit risk management practices. If the bank changes its watch list criteria, cross-selling policies, or top-up criteria, it is important to back test the results of the SICR criteria applied. Customers flagged into stage 2 may still be in the better segments from a business perspective, and therefore targeted by business for cross-selling and offering of top-ups at the existing rates. The SICR criteria may need to be revaluated, or business may need to refine its view of prime customers.
In the overall context of IFRS9, staging remains one of the more subjective areas within the expected credit losses (ECL) calculation. While it will never be 100% accurate, frequent monitoring and discussion with internal stakeholders will ensure that banks are able to best anticipate the impacts of their business decisions and the macroeconomic environment on the provisions, and will not be caught by surprise. Over time, as different stakeholders become more familiar with the methods and outcomes, the assessment of SICR will become more embedded in business practice, and therefore more aligned to bank strategy and process. In the meantime, banks should expect a lot more challenge and refinement in this area, and be proactive in understanding what is driving the result.