Reserving Methods8 min read

IBNR Estimation in Modern Practice: Beyond the Triangle

By Jonas Mohamed Osman Abdelghafour

Published 16 September 2026

How IBNR estimation is changing as insurers gain access to granular claims data, exposure information and near-real-time reporting.

Executive answer

IBNR remains the residual between an expected ultimate and what is currently known, but the information available to estimate it has expanded considerably. Reporting-delay distributions, exposure-based expectations and claim-level features now supplement aggregate triangles, particularly in classes where reporting patterns have shifted.

Where triangles still win

Aggregate methods remain strong where volume is high, patterns are stable and the audience needs a transparent audit trail. They also aggregate away a great deal of noise. The argument for granular methods is strongest where the triangle is thin, where mix has changed, or where the business needs answers at a segment level the triangle cannot support.

Practical hybrid approaches

A common and effective pattern is to estimate reported claim counts and average costs separately, use exposure-based expected loss ratios for immature periods, and reconcile the granular result back to a triangle-based control total. The reconciliation is what keeps the granular model honest.

Governance considerations

Granular methods raise data-lineage and reproducibility questions that aggregate methods do not. Version control of claim extracts, documented feature definitions and the ability to reproduce a prior quarter's number are prerequisites, not refinements.

Conclusion

The direction of travel is hybrid: granular where it adds information, aggregate where it adds stability, with an explicit reconciliation between the two.

Primary sources

IBNRgranular datareserving