Applying IFRS 9 expected credit loss (ECL) in a time of crisis

Severe economic downturns are modelled extensively but experienced far less often. This creates an inherent risk: when a true crisis emerges, it may differ materially from the scenarios institutions have prepared for. In such conditions, models and assumptions that appeared robust in testing can quickly become strained.

Since the Global Financial Crisis, periods of broad-based economic stress have been relatively infrequent. The pandemic represented a highly unusual and idiosyncratic shock, while subsequent years tested IFRS 9 expected credit loss (ECL) frameworks against a range of pressures, including supply-chain disruption, energy price volatility following Russia’s invasion of Ukraine, and the risk of global trade fragmentation. For much of this period, economic resilience has limited realised credit deterioration.

Recent developments, particularly affecting the Middle East, has increased macroeconomic uncertainty and the dispersion of plausible outcomes. For many firms in the GCC, the focus has shifted from abstract downside scenarios to assessing whether existing assumptions remain appropriate, and how ECL frameworks should respond as events evolve in real time. This article sets out practical considerations for applying IFRS 9 ECL in periods of acute stress.

When a crisis moves from theory to application

From a GCC perspective, a severe downside scenario may no longer appear remote. Conceptually, the response is clear: apply an existing downside or severe scenario within the ECL framework or draw on stress-testing outputs where concerns are elevated. In practice, several challenges quickly emerge.

1 – Scenario relevance

If an existing downside scenario does not reasonably resemble current developments, reliance on it risks anchoring decisions to an inappropriate narrative. Experience suggests that some of the most disruptive shocks — including the sharp inflationary surge of 2023 — were not fully embedded in pre-event scenarios. Revisiting scenario design when conditions change is therefore essential.

2 – Confidence in the economic response models (ERMs) that translate macroeconomic variables into credit risk outcomes

Periods of stress often expose limitations in these models, particularly where they were calibrated on data that did not include combinations of rapidly rising inflation, higher interest rates, and abrupt sector-specific disruption. In the current context, this includes potential discontinuities affecting oil trade, tourism, and cross-border capital flows.

3 – Macroeconomic data inevitably lags developments on the ground

Changes in borrower behaviour, liquidity pressure or counterparty risk are often visible within portfolios well before they are reflected in GDP releases or other national statistics. During these periods, bottom-up portfolio observations can provide an important complement to modelled outputs.

Why modelled signals can become unreliable under stress

ERMs are designed to inform judgement, not replace it. The macroeconomic data on which they rely is itself an approximation, constructed from surveys, revisions, and estimates. Under stress, models can become overly sensitive to a narrow set of drivers, over-fitting to historical relationships may become apparent, as can misalignment with the prevailing economic narrative.

There is evidence that some ECL models in the GCC are currently projecting stable or declining expected losses. In many cases, this reflects the strong role oil prices play within those models. Historically, higher oil prices have been associated with stronger regional growth, particularly where price increases were driven by demand-side expansion in the global economy.

The challenge arises when price dynamics are supply driven. In such circumstances, conditioning ECL estimates primarily on price levels can give a misleading signal. Economic outcomes depend not only on prices but on volumes and transmission mechanisms. If exports contract, investment slows or tourism activity weakens, elevated prices alone may not offset broader economic pressure.

GDP-based relationships are often more intuitive, given the well-established link between output and default behaviour. However, the structure of GCC economies has evolved significantly. Using the UAE as an example, while hydrocarbons continue to play a central role in fiscal revenues, non-oil sectors now account for a much larger share of economic activity. Models that remain tightly anchored to historical relationships may struggle to capture these structural shifts when conditions change rapidly.

Practical responses within IFRS 9 frameworks

A key initial step is assessing whether economic conditions have entered a severe scenario as defined within the ECL framework. Clear early-warning indicators, governance triggers, and escalation processes are critical in informing decisions around scenario selection, weighting, and management overlays.

As stress intensifies, post-model adjustments (PMAs) may be required to reflect risks that are insufficiently captured by benchmark models. For example, where exposures are concentrated in sectors such as tourism or trade-related industries, aggregate GDP-driven models may materially understate downside risk. Many ECL frameworks implicitly assume broad co-movement across sectors, an assumption that may not hold during periods of asymmetric shock.

Once a downturn is underway, reliance on top-down macroeconomic projections alone becomes increasingly limiting. Targeted analysis of sector-specific vulnerabilities, counterparty sensitivity and loss severity can provide timelier and decision-useful insight. Enhanced monitoring of large exposures and concentrations also becomes more valuable when official data is delayed or subject to revision.

Concluding observations

Historical distributions of past shocks can provide useful context when judging severity, but they also risk anchoring views to experiences that may not be representative of current conditions. The inflationary environment that emerged in 2022–2023 illustrated how risks can crystallise outside prevailing expectations.

In periods of genuine stress, macroeconomic forecasts inevitably play a supporting role rather than a definitive one. A clear understanding of portfolio composition, sectoral sensitivity and emerging behavioural signals becomes central to effective ECL estimation. Strong management information and governance frameworks allow firms to respond as conditions evolve, rather than relying solely on models that were calibrated for a different environment.

Want to learn more?

Send us an email if you would like to discuss ways your IFRS 9 ECL strategy can be optimised – info@4-most.co.uk.