Tanguy Touffut CEO of AXA Global Parametrics

How data science will help in responding to the next disaster

Environmental Challenges
Dec 1, 2017

Natural disasters cause major human and material losses and have become increasingly more frequent and severe. In the past year alone, a string of hurricanes tore through the Caribbean and the US, Mexico was struck by an earthquake, East Africa by drought and South Asia, parts of Africa and South America by landslide and flooding.

26 million people
fall into poverty due to natural disasters

on average every year

520 billion $
The cost of natural disasters on global economy

Source: World Bank

Natural disasters cost USD 520 bn in losses and force 26 million people into poverty every year, according to the World Bank. Climate change could exacerbate this even further, as disasters become more frequent and severe. With November’s COP23 climate summit and the Paris climate conference soon underway, these figures underscore the urgency for climate-smart solutions that better protect the most vulnerable.

Concrete effects of climate change on extreme events

Climate change is pushing up global temperatures but is also impacting extreme events.

The Intergovernmental Panel on Climate Change’s 2013 report found that globally, climate change is expected to cause hurricanes to occur less frequently. However, the strongest hurricanes, Category 4 or 5, are expected to become more frequent and the rain caused by hurricanes to increase as well. IPCC also highlighted that sea levels will rise due to climate change meaning that hurricanes will likely cause more flooding through storm surge.

One effective tool for building resilience to these shocks is insurance which injects capital into the economy for recovery efforts. Insurance already covers around 25% of losses from natural disasters globally. However, for low and middle-income countries, which are the worst hit and the least prepared, coverage is significantly lower and can drop below 10%. Insurance can and must go further.

Two major advancements could make this happen.

Innovation in data for speed

Firstly, weather data all over the world can now be accurately monitored thanks to technological progress in satellite and big data. Data quality has significantly improved and is much more granular. For example, daily rainfall levels can be tracked globally over just a few kilometers. And a long historic record of over 30 years of weather data is now available.

Weather data and monitoring has significantly improved in the past decades

Hurricane Irma struck the Caribbean and US this year as a Category 4 storm causing major damage. It was followed closely by Hurricane Jose as shown in this GIF based on NASA’s satellite data:

NASA satellite data is converted into a vegetation index measuring « greenness » over space and time. This animation shows 16-day and monthly vegetation coverage globally from 2000 to the end of 2013:

This has opened the door to parametric insurance, a new kind of risk transfer which uses satellite and other data as a proxy for losses on the ground. Payouts trigger automatically when the data indicates a disaster has struck. Satellite data detecting too little rainfall triggers payouts for drought-impacted farmers. Data on ground acceleration triggers payouts for those impacted by earthquake.

How parametric insurance works

Parametric insurance uses data that is correlated to a client’s damages or financial losses to create the most relevant cover. Weather data specifically is globally available so parametric insurance can be used in a range of weather-exposed industries, such as agriculture and energy, in both developed and developing countries.

While weather data is the most prevalently used today, parametric insurance can use non-weather data to cover other types of risks. For example, with fizzy, AXA offers parametric insurance covering flight delays.

Parametric insurance pays out quickly. It is more affordable than traditional insurance and can cover all parts of the world. This makes it ideal for financing emerging countries hit by catastrophes. Data-triggered payouts can reach a country within days of the event as no one needs to travel to the disaster-affected area to conduct a loss adjustment. The insurance also costs less because the data can be easily modelled and trusted by insurers. Countries can also use the weather data as an early warning tool to prepare for disaster when the data shows one is set to strike.

Data can act as a powerful early warning tool for disasters

Meteorologists use hurricane data to show potential tracks of a hurricane based on different models. The tracks in this spaghetti map were modelled for Hurricane Maria by the American Global Forecast System to predict where and when the storm was likely to strike. Other models go further in estimating the financial impact of or the cost of responding to a disaster. For example, the parametric drought model created by the African Risk Capacity predicts the worst to best case scenarios for a drought and the range of costs for responding to the affected population which is refined as the rainy season progresses.

Speedy payouts have major leverage benefits after disasters. Studies from Oxford University show that liquidity within weeks of a catastrophe is over four times more effective than funding many months later. Families which receive cash or food after disaster has struck do not have to sell their assets or pull their children out of school. Fast payouts therefore stop vulnerable people from falling into poverty traps.

The Vanuatu government received USD 1.9 mn in parametric insurance payouts from the Pacific Catastrophe Risk Assessment and Financing Initiative after the island was struck by Cyclone Pam in 2015.
Graham Crumb

Partnering with public and international sectors for scale

Besides data, another major advancement is that governments and international institutions have started understanding the value and importance of insurance, especially parametric insurance, in dealing with natural disasters and building resiliency.

The World Bank and the African Union have both set up regional risk pools for governments to parametrically insure their populations against disasters, which donors have funded. The G7 has pledged to insure up to 400 million more vulnerable people against climate risk by 2020. And this year, the UK and Germany set up a centre and a fund to provide technical assistance and fund climate insurance. The UN is also getting involved.

Ban Ki-Moon
Former Secretary-General of the United Nations

The UN Anticipate – Absorb – Reshape initiative is a multi-stakeholder partnership that focuses on accelerating climate resilience before 2020 for the most vulnerable by strengthening three elements: first, the capacity to better anticipate and act on climate hazards through early warning and early action. Second, the capacity to absorb shocks by increasing insurance and social protection coverage. Third, the capacity to adapt development to reduce risks at the national and international level.Investment in climate resilience can save 23,000 lives a year.

Developing countries are increasingly buying traditional as well as parametric disaster insurance or subsidizing premium for vulnerable individuals. The key advantage is scale. Governments’ existing structures can be leveraged to reach a large portion of people and to coordinate disaster relief funding.

Mexico received USD 150 mn in payouts from parametric insurance after being struck by a major earthquake in September 2017. The funds are being used for reconstruction and rehabilitation, helping hundreds of thousands of affected people. Caribbean islands struck by Hurricanes Maria and Irma received USD 54.4 mn in parametric insurance payouts within two weeks of the events.

These governments have proactively managed their disaster risk. They are less reliant on humanitarian aid and have not had to reallocate precious budget from key sectors such as health and education. Instead the private sector has been leveraged to shoulder some of the burden of disaster losses.

AXA is harnessing both breakthroughs for speed and scale

AXA is significantly investing in parametric insurance. This year we set up a dedicated subsidiary, AXA Global Parametrics, that is pushing innovation through data-based risk transfer. We are also focusing on refining data to ever more accurately represent actual losses on the ground.

Related content
AXA Global Parametrics: insurance, simplified
read more

AXA is deepening its relationships with governments and international institutions to create public-private partnerships, working closely with the World Bank Group and other such institutions. We strongly believe that partnering with the public sector and leveraging their know-how, distribution networks, and knowledge on the field, can aid us in reaching the most vulnerable.

We have continued to expand in this space since 2006 when AXA pioneered one of the first parametric deals covering the UN’s World Food Programme (WFP) against drought in Ethiopia. This year, we are participating, among a panel of reinsurers, in covering the Pacific Islands and the Philippines against earthquakes and typhoons losses of up to USD 220 mn. We are also insuring a number of African countries against drought losses including through the African Risk Capacity and the WFP’s R4 rural resilience initiative. These parametric deals have been facilitated by the African Union, the United Nations, the World Bank Group, regional development banks and other international institutions and donor countries. Working with such organizations has been invaluable to AXA in better understanding how to serve the public sector’s insurance needs in the face of climate change.

Using new approaches to insurance that leverage innovative technology, and working with the public sector, brings disruptive solutions to protecting people most exposed to disasters. Discussions at the Paris climate conference this month would do well to include sovereign parametric insurance as an effective tool for dealing with rising weather risk.

If we cannot control the extent to which climate change occurs, we can at least dampen its effects by harnessing data to protect more people and to protect better.