The ECB’s Paper on Novel Risks and IFRS 9 Overlays: A Missed Opportunity

Banks are continuously confronted with new risks that require evaluation. It is, however, quite difficult to handle emerging risks – particularly when attempting to estimate potential credit losses from these hazards through traditional models.

In a July 2024 paper, the European Central Bank (ECB) examines the challenge of capturing novel risks in loan loss provisions and offers its recommendations for best practices. But is the ECB’s guidance practical and logical?

Before we further scrutinize this issue, it’s important to first consider how banks forecast loan losses in compliance with the IFRS 9 accounting standard. Banks typically rely on models based on historical data to capture macroeconomic trends and their impact on key risk parameters of the loan book — probability of default (PD), loss-given default (LGD) and exposure-at-default (EAD).

Marco Folpmers

When calculating expected credit losses (ECL), banks apply macroeconomic forecasts to estimate and quantify future losses in the loan portfolio. However, the ECB argues that this conventional methodology is inadequate when dealing with “risks looming large on the horizon.”

This is where post-model adjustments (PMAs) come into the picture. IFRS 9 permits the inclusion of loss estimation for emerging risks through PMAs – also known as “overlays.”

The ECB’s paper cites the COVID-19 pandemic, geopolitical instability, climate change and inflation as examples of emerging risks that have challenged banks recently. While it is crucial to assess and reflect emerging risks in loan loss provisions (to ensure the effectiveness of forward-looking models like those that are used for IFRS 9 compliance), the ECB notes, ECL models have a hard time capturing these risks.

So, it is important for financial institutions to understand the pros and cons of PMAs and the circumstances under which they should be used. In their paper, the ECB states that PMAs are an adequate solution for loan loss estimation for emerging risks, especially for risks that do not have enough data to be modeled properly. But the central bank’s guidance on best practices for PMAs is not exactly clear.

Three Fallacies

The ECB’s analysis of PMA best practices, as well as their pros and cons, includes three questionable observations. Let’s now break down each of these fallacies.

Fallacy #1: Inflation is a novel risk. 

In their paper, the ECB identifies six emerging threats as novel risks: energy supply, supply chains in general, environmental risks, inflation, geopolitical risks, and the high interest rate environment. However, these risks do not all qualify as “novel.”

Inflation, for example, is not a new risk. There is ample historical data on inflation, including records of both low and high inflation periods, over the past 25 years (Figure 1).

Figure 1: Quarterly Eurozone Inflation

Source: ECB

As shown in Figure 1, banks can analyze time-series data and link them to observed default rates to estimate how inflation may impact defaults. In fact, many banks are already doing this using commonly employed techniques. The Vasicek model, for instance, incorporates a fitted correlation parameter to address the sensitivity of a risk parameter series to macro drivers. It is not only functional but can also be validated.

A review of economic history shows that inflation has been a persistent feature of societies for thousands of years, beginning with the transition from barter systems to monetary economies. For example, during the Roman Empire, inflation frequently occurred when grain supplies from Egypt were disrupted, causing prices to rise due to constant demand and reduced supply.

The famous 17th century “tulip mania” event, during which tulip prices soared dramatically, was one of the first documented speculative bubbles – and another historical example of inflation. Given this historical context, the ECB's inclusion of inflation — one of the oldest known financial risks — in its list of “novel risks” is perplexing.

Fallacy #2: The current batch of emerging risks is unprecedented.

The authors of the ECB paper imply that the times are changing and that emerging risks constitute a new problem. “This new risk environment poses a significant challenge to banking supervisors," the authors wrote. This language suggests that the current risk environment is unprecedented and requires adaptation from both risk managers and the ECB itself.

Contrary to this implication, financial risk managers understand that the broader environment in which they operate is perpetually in flux. Prior to COVID-19, of course, the financial world experienced market-altering events like the dotcom bubble in 2000, the collapse of Lehman Brothers (and the subsequent global financial crisis) in 2008, and the European Sovereign Debt Crisis of 2009-10.

During the latter fiasco, which saw soaring credit spreads for government bonds within the eurozone, a common risk management question was: "What is our exposure to Greece?"  In response to this question, risk managers adapted to new perspectives and made better use of cross sections of the existing data.

The point is that financial risk management inherently involves being vigilant, open-minded and responsive to new risks. This is no more relevant today than it has been in the past. Indeed, it is a fundamental aspect of the profession.

The risk environment is always evolving. It is an essential component of a forward-looking framework, by design. Risk managers must always adapt to new risks and anticipate unforeseen challenges. Therefore, the ECB's alarmist tone is unwarranted.

Fallacy #3: PMAs are potentially subject to earnings manipulation.

The ECB appears to suggest that PMAs are necessary but represent a second-best solution – a fallback if an emerging risk cannot be captured in modeled ECL.

In its somewhat ambiguous section on overlays, the ECB states that a decrease in PMAs for emerging risks has not yet materialized, but also conveys that, in the long run, "old" novel risks will gradually be absorbed into in-model ECLs. The reliance on PMAs should decrease, this reasoning suggests, as more data become available on novel risks.

The ECB further explains that when there is sufficient data for emerging risks, it is preferable to statistically model them than to subject them to PMAs. The supervisor argues that while PMAs can be useful, they can also be used for earnings manipulation. This concern is substantiated by the perceived positive correlation between pre-provisioning profit and the level of provisioning by PMAs.

However, this ECB analysis appears too simplistic. It lacks control variables and seems overly dependent on outliers. What’s more, in their paper, the ECB mentions the drivers of PMAs (including high interest rate regimes and shocks in the supply chain) — but does not take these into account in their analysis of potential earnings manipulation.

While it may be true that novel risks initially covered by PMAs should eventually transition to a traditional ECL model, the ECB’s analysis only scratches the surface and could benefit from greater clarity.

One can imagine a lifecycle model for novel risks and PMAs. In this lifecycle, new risks are identified, assessed, and initially quantified in PMAs, before being incorporated into in-model ECL. The lifecycle model would also establish criteria for transitioning through these stages.

Questions that merit further investigation include the following: What criteria should be used for new risks to be included in a PMA? Should this analysis be conducted for large sections of the portfolio (e.g., wholesale versus retail) or for more granular sections? When does a bank have sufficient data to consider transitioning a novel risk from a PMA to an in-model ECL? And under what conditions can PMAs be eliminated without transferring the risk to in-model ECL?

Only with a more detailed lifecycle model and clear stage transition probabilities can we address the question of whether PMAs will trend up, down, or remain stable in the long run.

Parting Thoughts

Addressing novel risks is an intriguing and essential aspect of risk assessment, model updates and overlays. Given that the IFRS 9 framework is inherently forward-looking, understanding and mitigating these novel risks is particularly crucial for the effective application of this accounting standard.

Regrettably, the new ECB guidance lacks a well-defined framework for the lifecycle of PMAs. Moreover, the supervisor does not adequately substantiate its serious allegation of earnings manipulation through PMAs.

Overall, the ECB’s recommendations fall short in providing the background and methodologies needed to structurally enhance banks’ existing frameworks for forecasting credit losses. While the guidance rightly emphasizes the need for PMAs to address these emerging risks, it fails to offer feasible, clear recommendations for best practices.

Dr. Marco Folpmers (FRM) is a partner for Financial Risk Management at Deloitte the Netherlands.