The property data and analytics company CoreLogic® recently released its updated Japan Earthquake Model on Nasdaq’s catastrophe risk modelling service. This is CoreLogic’s second model to be made available through the Oasis Loss Modelling Framework (LMF) and Nasdaq, following the launch of its U.S. Earthquake Model in February 2020.
We sat down with David Gregory, product manager for the global earthquake product suite at CoreLogic, to discuss the new model and latest developments within the catastrophe risk management department in CoreLogic.
As a bit of background, tell us about CoreLogic and your mission.
CoreLogic is a property data and analytics company with a strong division in insurance solutions, including catastrophe model development. We support financial institutions, real estate professionals, government agencies, insurance carriers and more in the pursuit of enabling millions of people to find, buy and protect the homes and businesses they love. This includes helping protect from risks such as natural disasters. This comprehensive, end-to-end reach across the property ecosystem enables us to deliver unique data-driven insights to more than 500 insurance carriers in the U.S. in addition to those in the Lloyd’s Market and internationally.
How does catastrophe modelling fit into CoreLogic?
I joined CoreLogic as part of its acquisition of EQECAT back in 2013. Since then, what has really made my role interesting is the broad access to the wider data repository and expertise that CoreLogic has and working to identify and develop new ways to increase the robustness of the earthquake models using this data.
We talk a lot about loss validation in cat modelling and how to best calibrate models with claims data. However, without the extended knowledge of the property makeup of buildings, claims data is of limited use.
With the CoreLogic property characteristics database, we know not just the structure type, year built, etc. of buildings, but even have information on replacement cost and secondary characteristics—such as whether a building has cripple walls or base isolation. These data points are invaluable for earthquake modelling. For example, in the 2011 Tohoku, Japan, earthquake, we used our in-depth data analytics that helped us better quantify and correlate claims data to better understand the risk. We maintain this data in the U.S., New Zealand, and Australia, and we have a data partner in Japan.
What drove you to develop your new model in the Oasis format?
Japan earthquake risk is regarded as one of the five traditional key catastrophe perils, so it made sense for us to make it available in the Oasis LMF and thereby more easily accessible to a wider audience, for example through the Nasdaq Risk Modelling for Catastrophes service. The model was updated at the start of 2020 to the latest available science. In terms of capturing Japan earthquake risk, there is typically a lot more discussion on how to capture the hazard than compared to say, the U.S.
How does CoreLogic approach this complexity in the design of your model?
The Japan government agency model (ERC-HERP implemented by the NIED), which is used for the Japan earthquake building code, is often considered somewhat conservative when it comes to several components of the model. When it comes to protecting human lives in relation to the building code, being conservative isn’t the least sensible thing to do. However, when it comes to establishing a commercial earthquake model for insurance pricing, it was clear that a different approach was needed.
Our earthquake scientists took a fresh look at the risk, with a particular focus on where it matters for our clients’ risk, in how we capture the risk on the Nankai, the seismic source that drives over half of the insured risk in Japan. This was made possible by using our 300,000-year simulation, where we have managed to create a robust model that captures the risk while reducing the epistemic uncertainty in this model update.
Did you face any unique issues in the conversion of Japan Earthquake on Oasis compared to when converting the U.S. Earthquake model?
This is a complex process, and while I’m sure that the team encountered many challenges along the way, the first thing that springs to mind is the unique financial conditions that apply to certain Japanese insurance business, more typically for the mutual companies.
The team at CoreLogic worked together with Oasis, Nasdaq and our clients to implement these conditions in the system and ensure they are configured to work for our model. Cross-organizational projects like this add complexity as we work on different systems and work to different rhythms, but so far, it has worked seamlessly.
After the release of Japan earthquake, what’s next?
We find ourselves in a new wave of cat risk management, where flexibility is key, and at CoreLogic, we are committed to making our analytics more readily available in our clients’ existing workflows.
As such, we have two development tracks in our roadmap, one where our scientists continue to work on peril region model updates and another that focuses on creating software products in a platform-agnostic framework to enable models to be made available both on our own platform RQE® and through the Oasis LMF and via additional providers, such as Nasdaq Risk Modelling for Catastrophes.
As for what’s next, we have over 180 peril region models, and we are currently encouraging marketing feedback to define the next model migration—we are hearing the strength of our European Windstorm and U.S. Wildfire models would make good candidates, but we welcome your thoughts.
To learn more about CoreLogic catastrophe risk modeling, visit their website.
Curious about CoreLogic’s Japan Earthquake Model? If you are a re/insurance firm or broker, you can register your interest to evaluate the model on the Nasdaq Risk Modelling for Catastrophes service.