IFRS 9 isn't just a regulatory burden—it can be a powerful tool for understanding customer behaviour, driving smarter credit decisions, and enhancing profitability when leveraged beyond compliance.
IFRS 9’s Expected Credit Loss framework can do far more than fill a reporting binder. That was the message from a late session at CeFPro’s annual Risk Evolve conference this year.
Jakob Lavröd, Senior Quantitative Risk Analyst at Handelsbanken in Stockholm, argued that banks are leaving value on the table by treating ECL as a box-ticking exercise rather than a strategic engine.
“Stop aiming for adequate compliance and start extracting decision-useful signal,” Lavröd said.
He framed the discussion in a long view of conservative credit culture, recalling how Handelsbanken’s history was shaped by a fight over loss provisioning during Sweden’s 19th-century railway boom.
That heritage, he suggested, should be repurposed for modern IFRS 9.
“Sound risk management is not a regulatory burden - it is the defining function of a bank,” Lavröd said, pointing to post-SVB supervisory soul-searching as proof that resilience beats box-checking.
The hinge, he argued, is Probability of Default. Instead of starting with 12-month PD for Stage 1 and switching to lifetime PD in Stage 2, his advice was to start with lifetime expectations for every exposure at origination.
That, he told delegates, better matches real credit decisions and turns IFRS 9 into a dynamic updater - a way to measure deviations from the initial baseline rather than resetting the baseline every reporting date.
“Think in lifetime terms on day one, then update continuously as evidence arrives,” Lavröd said.
To do that credibly, Lavröd urged banks to adopt survival modeling. By modeling hazard rates and full default-time distributions, institutions can re-estimate risk as conditions change - without waiting for a 12-month cycle to close.
He cited a COVID-era case in unsecured lending where the survival model detected a counterintuitive fall in risk as discretionary spending collapsed and government support surged.
“The data said risk went down, so the right answer was a negative ECL adjustment - and yes, that was controversial,” Lavröd said.
Exposure at Default received similarly unsentimental treatment. In Sweden’s broker-heavy unsecured market, upfront commissions capitalized as transaction costs under IFRS 9 can warp effective interest rates if customer attrition is high. That makes EAD behavioral.
