Skip to main content

Designing liability scenarios

This report presents an approach developed by Arium which combines insights provided by scenarios and supply chains to quantify losses that could arise from liability events.

30 Nov 2015

Designing liability scenarios

This report touches upon the potential to develop a stochastic modelling capability for liability exposures, and proposes mechanisms by which this could be achieved.


Liability exposure management has historically been a challenge for insurers. Liabilities often stem from a complex interaction of legal and socio-economic factors which can make this kind of problem hard to represent and the exposures hard to capture in a form that lends itself to systematic study. Lloyd’s is investigating different methods that aim to reduce uncertainty in this area, and this report presents an approach developed by Arium which seeks to harness the power of scenario design and supply chains.

Arium’s framework combines a scenario of a product or service causing harm with publicly available trade data to build a map of the spread of potential liability through industries. This map is filtered using information about the harm caused and policy information, and used to calculate an aggregate loss for the scenario. The scenario may then be split into sub-scenarios, and parameterised to calculate a per policy loss. The report also sets out case studies of both a financial and a non-financial liability scenario developed using this framework.

Key facts:

  • Mapping supply and distribution chains illustrates the fact that liability can arise or end up in industries beyond that from which a potentially harmful product or service originates.
  • Scenarios can be developed using data from a diverse range of sources both within and outside an organisation. While this collaborative approach is resource-intensive, it can enable more thorough design and stress testing of scenarios.
  • For liability scenarios and the resulting loss estimates to provide meaningful insights for insurance practitioners, portfolio data should be captured in a format that is optimised for analysis. Appending basic corporate information to policy data is necessary to run scenarios and accumulations, and is likely to result in a more consistent and comprehensive dataset than most insurers currently possess.
  • Despite continuous progress in modelling capabilities for property exposure to natural catastrophes, modelling for liability has remained a challenge for insurers. Advances in the quantification of liability scenarios are a promising step towards improving insurers’ understanding of emerging liability risks.