Skip to main content

Earthquake science for (re)insurance decision-makers

Medioambiental|Reinsurance|Willis Research Network
Climate Risk and Resilience

By Myrto Papaspiliou , Massimiliano Arizzi and Crescenzo Petrone | November 9, 2020

Advanced analytical solutions are increasingly sought in the reinsurance industry and the Global Earthquake Model can be leveraged to improve capital management and control earnings volatility.

The Global Earthquake Model (GEM) foundation is a key Willis Research Network (WRN) partner with whom we have collaborated on a number of initiatives over the last few years to deliver earthquake risk assessment solutions for Willis Towers Watson clients. GEM is a world leader in earthquake science and has developed a global, fully probabilistic, earthquake hazard and risk model that is continuously being updated.

On 23 September Willis Towers Watson held a webinar looking at earthquake science for (re)insurance decision-making, focusing on how the industry can leverage the latest research outputs from GEM. The event considered two main themes: using GEM’s datasets to enhance loss assessment and applying their probabilistic models for novel and rapid loss estimation. The event featured a presentation by GEM scientists and Willis Re earthquake experts on the scientific developments and their translation into (re)insurance business applications. The developed solutions can be fundamental in assisting clients with a number of natural catastrophe risk management themes:

  • Regulatory requirements are creating a need to demonstrate a view of risk in line with one’s exposure profile as well as an understanding of model limitations and their materiality.
  • Alternate views of catastrophe risk are increasingly sought in developing a robust loss estimation and defining economic capital needs.
  • The ever evolving (re)insurance market conditions make technical discussions during placement more relevant.
  • Innovative research and products are needed to address unmodelled risk and the sound determination of loss size and frequency of events to mitigate earnings volatility.

Enhancing loss assessment using GEM exposure and vulnerability databases

The first session of the webinar focused on the advances in the vulnerability and global exposure models developed by GEM, which are extensively used by Willis Towers Watson to develop the Willis Re View of Catastrophe Risk worldwide. Key takeaways of the session were:

  • The use of GEM exposure data can refine the view of risk and support robust, technical broking arguments in the reinsurance placement, particularly in regions and for portfolios with ‘unknown’ construction typology or other primary modifiers.
  • The GEM vulnerability database provides a key role in the comprehensive approach towards risk quantification developed at Willis Re and depicted below, allowing us to create a customised view of risk suitable to individual portfolios and underpinned by a transparent and scientifically robust methodology that supports reinsurance placement and improves our confidence in defining capital needs.
The diagram shows the process of how GEM hazard data is used:
Figure 1. Leveraging GEM vulnerability database in loss modelling: a comprehensive approach towards risk quantification.
  1. GEM Hazard Data collected
  2. Location EP Curve: combining hazard with vulnerability from Willis Re vulnerability framework.
  3. Probabilistic loss approximation: spatially correlated aggregation adapted from KatRisk Solokat+. Building an ELT from loss layers
  4. Aggregated Portfolio EP curve: alternative view of risk based on GEM models with limited computational effort

Leveraging GEM model for probabilistic loss estimation: two alternative approaches

The focus then moved on to the updates to the global mosaic of 30 models behind the GEM risk model, with a discussion of the new key hazard modelling features currently being implemented in regions like China. These models can help decision making through:

  • The development of stochastic event sets from GEM models that allow calibration of existing models for use in decisions around reinsurance.
  • Fully probabilistic models developed with the OpenQuake engine from GEM, such as the Middle East and North Africa earthquake model. The availability of such model enabled Willis Re to define and calibrate a parametric cover with minimal basis risk for all residential buildings in a country in North Africa, a very relevant risk transfer solution, particularly for emerging markets.
  • An exposure driven, reproducible and robust method to rapidly produce alternatives to vendor views of risk, without the need to run a computationally expensive, fully probabilistic, risk model. Willis Re has been developing an innovative solution over the last few months to provide much needed support for reinsurance placement and in defining capital requirements in unmodelled territories or areas with outdated vendor offering. The low computational effort of the bespoke methodology shown below, that still leverages the latest science by GEM, is at the forefront of the innovative Willis Re service offering available to its clients.
The diagram shows how Willis Re manage risks:
Figure 2. Innovative portfolio modelling for reinsurance decision-making: rapid conversion of GEM model outputs to portfolio Exceedance Probability (EP) curves.
  1. Evaluate models
    • Compare: Modelled vulnerability curves with relevant GEM database curves.
    • Stress test: Assumptions and model relativities using extensive GEM data.
  2. Build solutions where needed
    • Adjust: Modelled vulnerability to align with empirical mean view.
    • Validate: With observed damage data from recent events.
  3. Customise to own view of risk
    • Enhance: Model robustness and representation of risk.
    • Impact: Provide a transparent and tailored view of risk.

The increasing role of analytics and the input from academia in managing risk under current market conditions was clear throughout the seminar; leveraging GEM outputs in business decision-making can significantly improve capital management and control earnings volatility.

The recording of the webinar is available on request.

For more information on any of the topics mentioned please contact us by email.


Lead Earthquake Specialist
Willis Re International

Myrto is an Earthquake Specialist in the Model Research & Evaluation team for Willis Re International. The team is in charge of developing the Willis Re view of risk by assessing, comparing and adjusting catastrophe model vendors for all perils/territories in the international business. Myrto has been responsible for the co-ordination and delivery of earthquake catastrophe models evaluations for more than 30 territories around the world. Her areas of focus include technical evaluations, supporting clients in developing their view of risk and communicating study outputs with reinsurers. Myrto holds a PhD on Earthquake Engineering and prior to joining Willis in 2012 she worked as an earthquake engineer in a leading engineering consulting company.

Executive Director, Head of EMEA WS London Team
Willis Re International

Massimiliano Arizzi joined Willis Re in 2005 as a qualified actuary in the Italian office. In 2013 he moved to the London HQ where he now leads the “EMEA WS London”, a central team with commercial and analytical expertise to promote service offering excellence fully integrated with the regional offices. Member of the management board for the Region, he is involved in the handling of large domestic and international clients and has gained substantial experience on traditional Property, Casualty and Financial lines as well as on structured reinsurance and capital management projects.

Head of Southern Europe Catastrophe Analytics

An earthquake specialist at Willis Re, Crescenzo develops risk solutions to help clients accurately understand and mitigate their own risk.

Related Capabilities

Contact Us