About The Hidden Grammar of Reserving: A Probabilistic Foundation for the Chain-Ladder Method with Claim Counts (14/12/2026) - 1 CPD

The Hidden Grammar of Reserving: A Probabilistic Foundation for the Chain-Ladder Method with Claim Counts

The Chain-Ladder method has dominated non-life reserving practice for decades, yet it remains an unusual statistical object: despite its widespread use, it lacks a full probabilistic foundation. While stochastic extensions such as the Mack model and the Over-Dispersed Poisson framework have provided important advances, fundamental questions remain about what the method is actually estimating, how uncertainty should be interpreted, and what information a run-off triangle can truly reveal.

In this session, Robin Van Oirbeek will present the Negative Binomial Chain-Ladder model, a full-likelihood extension of the classical Chain-Ladder that is derived from an underlying claims-generating process and provides a structural interpretation of overdispersion.

More broadly, Robin will use this model as an entry point into a wider research programme called ‘The Hidden Grammar of Reserving’. The central idea is that behind classical reserving methods lies an often implicit structure governing what can be learned from run-off triangles, what assumptions are being made, and which quantities are fundamentally identifiable.

Rather than viewing reserving techniques as a collection of disconnected tools, the talk will show how concepts such as prediction, uncertainty, credibility, overdispersion, identifiability and model assumptions fit together within a common framework. The aim is not only to explain a new reserving model, but also to provide a different perspective on how classical reserving methods relate to one another and what can realistically be learned from aggregate claims data.

The Negative Binomial Chain-Ladder model received the 2026 CAS Ronald Bornhuetter Loss Reserve Prize and serves here as an entry point into the broader Hidden Grammar of Reserving research programme.

After this session, you will: 

• Understand how the Chain-Ladder method can be embedded within a full probabilistic framework.
• Gain intuition for different sources of reserving uncertainty and how they arise.
• Develop a clearer understanding of what run-off triangles can and cannot tell us.
• See how seemingly different reserving approaches can be viewed as parts of a broader conceptual structure.
• Leave with a fresh perspective on the foundations of modern loss reserving.

Practical information
The session will be held in English and takes place online via Microsoft Teams. The link will be shared two days in advance.