Mind (the data) gap, with Jon Gascoigne, Willis Towers Watson

Sept. 5, 2018

In this series of blog posts, we talk with leading figures in the insurance industry about identifying and closing the data gap. Today I’m joined in conversation with Jonathon Gascoigne, a Senior Risk Adviser, Capital, Science & Policy Practice WTW

JG = Jon Gascoigne

JC = Jim Craig

Jon, can you start by giving your thoughts on the various initiatives around data that you can see at the moment please?

Of course, the significant tropical cyclone events that took place across the Caribbean in Q3 of 2017 both focus the mind with event response reporting for clients as well as providing further extreme event examples to test and refine existing models. Such events also highlight the ranges of data availability and quality that exist by territory. How can most effectively fill such data gaps and also access information of hazard and exposed value at risk that may be scattered around from varied sources?

The work underway with Oasis Loss Modelling Framework (LMF) to open source the modelling tools and the resultant models is particularly important and something we'd like to unpick further. Similarly, the on-going work sponsored by the Insurance Development Forum (IDF) to signpost all the modelling tools and providers through the CatRisk Tools website provides support to catastrophe modellers in understanding physical risk, fundamentally in terms of frequency, severity and correlation

Thanks Jon. Obviously, physical events are drivers to stimulate action, what else do you see is happening more proactively?

There’s increasing technical dialogue between the re/insurance industry and the public sector and international organisations based on assessing and enhancing societal resilience. Emergency response to devastating events such as Hurricane Irma put this work to the test, highlighting the crucial time-dependent aspect, with CCRIF (formerly Caribbean Catastrophe Risk Insurance Facility) providing parametric insurance payouts at sovereign level with 14 days. The on-the-ground logistical work provided by such non-governmental organisations like MapAction is also invaluable here and hopefully ever-evolving technological developments which promote greater co-ordination of limited resources at such times. Getting a MapAction button on OasisHub could be a part of this!

Regarding ‘natural capital’ and valuing nature explicitly, rather than the more familiar built environment, exciting work is taking place in applying insurance sector risk modelling and possibly financial products to the ecosystem services that biodiversity provides. What gets measured gets managed. And the chance for another Oasis Hub data button? This is still a relatively niche field and so it wouldn’t get pressed so often at the moment.

Ok, I hear you, we’ve had an initial conversation with some of the people at MapAction.

We are hearing “climate change data” more and more from a wider range of people. What’s the interest from an insurance perspective?

Insurers clearly have a vested interest in seeing the impact of climate change mitigated, however at an immediate level there’s a time horizon issue of insurance policies generally being annual and so analytics must be geared to the risk profile of that policy period. However, understanding ‘non-stationary’ signals is data, be that ENSO, NAO or climate trends helps interpret historical data. And it should be noted that to contextualise the sensitivities on any climate change impact, you have to be able to compare this with the baseline risk profile of a portfolio or location.

There is an opportunity to use insurance not only as financial compensation after an event but also as a mechanism for supporting risk reduction, which is in both the industry’s and society’s interests. ClimateWise was established in 2007 by the insurance industry to better communicate, disclose and respond to the risks and opportunities associated with the climate-risk protection gap - the growing divide between economic and insured losses.

Considering physical climate risk, the broader financial sector of banking and investment is looking to the insurance industry’s experience of quantifying weather-related hazards to inform and assess impact on corporate performance, as well as unlocking resilient investment opportunities globally. The Task Force on Climate-Related Financial Disclosures (TCFD) has provided a remarkable stimulus to this activity. This requires consideration of chronic risks such as water and heat stress, as well as acute extreme events. Crucially, catastrophe models provide a strong analytical foundation as probabilistic modelling is implicitly forward-looking by using historical data to construct synthetic sets of many thousands of years of events; this process easily incorporating ‘climate conditioning’ with IPCC scenarios, for example. I’d also like to mention lively recent work from legal and banking directions in this arena from Clyde and Co and EBRD.

Climatic impacts are more damaging in emerging economies which have increased vulnerability. Extending the societal resilience opportunities of re/insurance risk transfer will require support from data analytics and models. Expanding the community of model providers could dramatically help in providing this coverage, so long as consistent metrics are applied and issues of transparency and interoperability are dealt with, incorporating the Oasis framework, for example.

What would a list of missing data, relevant to the projects you are working on look like?

In developing new markets and business opportunities, much of my work engages with sectors and threats beyond the traditional day job of catastrophe modelling in re/insurance, where data requirements are better understood. So personally, with this in mind I’ll mention a few data needs here that push this envelope of the familiar:

Natural hazards data for most of Africa and Asia. What geographical and temporal resolutions are available? Also consideration of ‘slow-onset’ hazards such as drought.
Utilities and critical infrastructure, the roads and bridges? What is their level of build and maintenance quality? How to quantify systemic risk fro regions?
Education sector data, by that I mean where are the schools? How big? What type? How many students attend? Systematic analysis of the relationship between humanitarian and development needs would greatly help improved co-ordination.
The “blue economy”: vulnerability characteristics of species and matrices of coral reefs, seagrass and mangroves to storm surge, pollution and temperature change? Can we assess damage and reinstatement to such ecosystems themselves as well as to human activities they may support and protect?
Application of climate change data – what scenarios are most appropriate and how to apply then? How to develop plausible and appropriate stress tests?

Then we also need to consider the meta-data issues:

How extensive are data-sets? How complete and reliable is the data?
Is much required data already out there somewhere? If so what is the time and resource required to find it, collate it, and curate it? Licencing issues?
Commensurable? Are we able to compare ‘apples with apples’?

Wow! Thanks Jon, that’s quite a list.

Are there any other considerations for risk data-gathering you’d like to mention?

The city-based focus is representing an ever-increasing lens through which to assess risk, as they represent such concentrations of exposure – with increasing urbanisation trends, particularly in hazard-prone areas – but which may also be able to exercise some autonomy at the meso-level over implementing disaster risk reduction and resilience measures. I’m reminded of a recent Lloyd’s of London / ARUP report focussing on city infrastructure with risk case studies emphasises the need for integration of both physical and systemic risk considerations for service continuity. The Lloyd’s Cities@Risk Index, which is shortly to be updated, provides a rich quantified picture across geographies and threat classes beyond ‘nat cat’, and applies the innovative scenario-based work of the Cambridge Centre for Risk Studies, which enables such a systematic and synthesising view.

Thank you, Jon. If people would like to give feedback please add a comment below. If you’d like help in closing the data gap please get in touch at hello@oasishub.co.

Useful links

http://www.swissre.com/media/news_releases/nr20180410_sigma_global_insured_loses_highest_ever.html

https://oasislmf.org/

https://oasishub.co/

https://catrisktools.oasishub.co/

https://www.ccrif.org/

https://mapaction.org/

https://www.cisl.cam.ac.uk/business-action/sustainable-finance/climatewise

https://www.ipcc.ch/pdf/assessment-report/ar5/syr/SYR_AR5_FINAL_full.pdf

https://resilience.clydeco.com/

http://www.ebrd.com/news/2018/ebrd-gceca-publish-recommendations-on-physical-climate-impacts-in-financial-reporting-.html

https://www.lloyds.com/news-and-risk-insight/risk-reports/library/society-and-security/arup