Natural catastrophe losses continue to grow due in large part to the effects of climate change.
A record-setting 2020 Atlantic hurricane season, with 30 named storms, 12 of which made landfall, contributed to $97 billion in global insured losses – 40% above average since 2000. Tropical cyclones and flooding continue to be responsible, in the aggregate, for the largest portion of global economic losses caused by natural disaster events.
Increasing uncertainty and higher frequency and severity of weather catastrophes over the past 20 years are putting many companies to the test, with risk managers under pressure to prepare for an increasing number of catastrophic perils, and to account for them in their risk management strategy.
A record-setting 2020 Atlantic hurricane season contributed to $97 billion in global insured losses.
With the rise in catastrophic events and losses over the past 25 years, financial analysis of catastrophe events known as CAT modeling has taken on an increased level of importance. CAT modeling is used primarily for insurers and reinsurers to evaluate and manage catastrophe risk from many different types of catastrophes, ranging from hurricanes and flood to earthquakes and wildfires. CAT modeling has taken tremendous steps in the past quarter century. More complex algorithms and new data sets have enabled additional perils to be modeled while increased computing power supports more detailed modeling at finer resolutions.
How CAT Modeling Works
Models are created by teams of cross-disciplinary professionals, who work together to create software, data and algorithms to help simulate the potential impact that catastrophes can bring in specified circumstances. Developing sophisticated CAT models can be complex, with the needed expertise from variety of technical disciplines, including meteorologists, seismologists, geologists, engineers, mathematicians, actuaries, scientists and statisticians. Models also use policy and financial data from insurers and reinsurers, including coverage values, deductibles, limits, and other policy-level information, to create a probability of loss profile from different event scenarios exceeding certain levels or other parameters.
For insurers and reinsurers, CAT modeling analyses can cover a wide swath of territory (e.g., states, regions, countries) to help them make critical risk decisions on their book of business. Model outputs often provide underwriters with a better way to analyze and evaluate the overall level of capital that may be required to pay potential claims and to potentially allocate capital between individual risks within their portfolio. Models can also help with ratemaking decisions, business appetite, and the ability to refine reinsurance decisions.
CAT modeling works on the micro level as well. One of the greatest concerns of risk managers is the unknown, or surprise events that are difficult to predict or evaluate in advance. Risk managers seeking to take control of their risk should be aware of the potential for catastrophic losses in their locations, including, for example, hurricanes and flooding in the southeast and eastern seaboard and earthquake exposure on the Pacific Rim, and many risk managers work with their brokers to take action to control for the unknown and focus on building risk resilience. While natural catastrophes bring volatility to many businesses, they can be managed, and the unknown can be better accounted for with data and analytics based in part on CAT modeling.
Further, as climate change becomes more acute, the accuracy of its impact in CAT modeling is becoming critical. Climate change factors can be layered on top of the base CAT modeling analysis to help risk managers gain a more accurate loss estimate. CAT models also help risk managers manage their event response and mitigate loss by having claim adjusters on the scene as soon as possible.
Through the use of modeling information received, risk managers can make important risk decisions about their risk management framework, which may include the use of captives or parametric insurance, which are increasingly helping risk managers rethink their access to capital by providing a complementary risk solution to their traditional risk transfer program with additional resiliency and capacity. Parametric insurance, for example, can be placed on specific exposures and paid out when a triggering event occurs.
Here’s a deeper dive on both hurricane/flood and earthquake modeling:
Windstorm and Flood Modeling
In 2020, four hurricanes were on the list of top 10 global insured loss events – and two were in the top three. Of the top 10 global economic losses in 2020, three flood events made the list and seasonal floods in China was number one on the list.
Hurricane models have grown in importance following major events like Hurricane Katrina in 2005, although the earliest modeling submissions to the Florida Commission on Hurricane Loss Projection Methodology were made in the mid to late 1990s, following Hurricane Andrew in 1992.
The impact of climate change on the frequency and severity of named storms has also made accurate modeling even more important to insurers, reinsurers and risk managers with exposure (e.g., insured properties) in hurricane-prone areas. Enhancements made to windstorm and flood modeling over time have provided a more accurate risk view. Breadth of peril selection has improved to provide a more accurate exposure view as well. Models can now provide inland and coastal flood loss estimates and have also improved on the correlation of losses from hurricane and flood, as the two exposures are often related to each other. The ability to link multiple perils is critical for an accurate estimate of insured loss.
The risk manager is better able to refine loss estimates for a specific property by accounting for terrain conditions, impact of wind conditions, building specifications, and other regional or property-specific characteristics. Enhanced modeling also accounts for building attributes, including the number of stories and construction features, and gives risk managers the ability to better understand and refine loss estimates for risks of different types.
By working with a broker, risk managers can often use CAT modeling data and analysis to rethink their access to capital – for instance, by looking at current risk transfer options, the prospect of retaining risk on their balance sheets, transferring to the traditional (re)insurance market, or looking at alternative structures such as various captive options or parametric insurance.
It wasn’t too long ago that insurers would keep track of their earthquake exposures by sticking push pins into a map. Our fast-paced business world today requires a much more refined approach to earthquake modeling, which has become more sophisticated and detailed in recent years.
While advancements in CAT modeling technology has allowed insurance carriers and reinsurers to gain a better and more accurate view of earthquake exposure based on loss estimates in earthquake-prone areas, risk managers can also work with their brokers to better understand risk exposures site-specific loss estimates, based on soil conditions, construction, and other regional or property-specific characteristics.
Earthquake activity does not show large fluctuations on an annual basis, nor does weather directly impact quake activity, but the peril’s likelihood of striking in earthquake-prone areas is a major concern to businesses located there.
It is nearly impossible to be precise on where the next “big one” will hit. Instead, models take a stochastic view of risk using science to support data-driven decisions. They account for a wide range of possible scenarios to assist risk managers with understanding where losses could happen, and the magnitude and frequency of the financial impact of potential losses should an earthquake occur. The use of better data is critical for risk managers to understand their organization’s risk exposure and to make more informed risk transfer decisions, including the possible use of captives or parametric insurance.
It is nearly impossible to be precise on where the next "big one" will hit.
By working with their broker to conduct CAT modeling and other financial analyses or discussions, risk managers are also better positioned to refine loss estimates for a specific property, which will provide the business with a sound understanding of earthquake-prone locations, construction planning, cost of insurance, and other potentially relevant risk management factors.