激化する大災害の保険業界への影響

It’s spring here in northern California. The trees are blooming; the bulbs are up; the roses have budded out. My garden is looking particularly spectacular right now. This has been the one silver lining in a pandemic world - being home to enjoy the garden. But, I’m not sure how long this will last. I read this week that although we usually get 24 inches of rain a year, my town has only had nine. And the year is almost over. That means we’re likely to have water rationing later on and my garden will suffer. However, a suffering garden is not the only impact of a low rain year. Five of California’s six largest wildfires on record happened last year. Nine of the ten biggest wildfires happened in the last ten years. And this is our second year of drought. (As a side note, in California, many of the residents already had masks when the pandemic began because the smoke and ash from the fires makes it difficult to breathe without them.)

We’re not the only state looking ahead to potentially increased disasters. With nearly two-thirds of the United States abnormally dry or worse, the National Oceanic and Atmospheric Administration’s official spring outlook predicts a prolonged and widespread drought for a large swath of the country from Louisiana to Oregon. The governor of North Dakota has just declared a drought disaster and wildfires have already begun to increase.

It’s more than droughts and wildfires that we’re facing. The U.S. National Oceanic and Atmospheric Administration (NOAA) has increased the number of storms that they consider an average Atlantic hurricane season from 12 to 14. The prior period-of-record, 1981 – 2010 showed on average 12 named storms, 6 hurricanes and 3 major hurricanes. The average over the thirty year period of 1991 – 2020 is now 14 named storms with 7 hurricanes. And on Thursday, Colorado State University forecasters said they expect 17 named storms and eight hurricanes, four of which will be major, in the 2021 season.

Of course, the insurance industry is well aware of these happenings. Chief Risk Officers run the models regularly to assess the potential impact on their books and determine what actions they can take. But looking forward is always a challenge. After all, the weather could go back to historical norms, or could continue to change – either continuing along the current trajectory, radically changing, or bouncing back and forth wildly. The data needed to generate relevant predictions is often out of date, incomplete, or inaccurate. While multiple models exist, models are still missing for some key areas and don’t always match a carrier’s actual book of business. And technology hasn’t moved forward quickly enough. Running a scenario is slow – roll ups take a long time. There is still a lack of real time guidance to underwriters on a risk by risk basis.

The result is sub-optimal exposure management and cross-geography optimization which results in an inefficient use of capital. Insurers are forced to take broad actions instead of surgical strikes, and this can result in a higher cost of reinsurance.

So where can insurers take action? And what kinds of actions can they take? Potential interventions fall in a couple of areas:

• Individual account underwriting

– Insurers can establish or modify existing underwriting guidelines regarding the kinds of exposures they’re willing to take and the locations of those exposures. They can use analytics to identify those risk characteristics that have the biggest impact on susceptibility. They can analyze policy structures or develop new pricing mechanisms. After all hurricane deductibles and by-peril rating are relatively new mechanisms. This has implications for the policy administration system. Insurers may need the ability to easily change underwriting rules or rating plans.

• Book /territory management

– Insurers can use data and analytics to more accurately determine risk drivers across their book of business to modify appetites or underwriting guidelines. They can evaluate capital adequacy and allocate capital across territories – looking at balancing their book of business. This requires a greater use of data than typically found in a policy admin system - and to provide guidance to an underwriter at the point of sale, means the policy admin system needs to be able to monitor the overall book of business and fire off underwriting rules when overall capacity hits a predetermined point.

• Capital management

– With potentially increasing catastrophes, insurers may need to reassess their own reinsurance needs. And of course, it’s not just natural disasters. New exposures continue to lurk – whether pandemic related, or cyber related, or based on new technologies such as nanotechnology or driverless cars. Insurers are increasingly looking to structure sophisticated risk transfer programs using reinsurance, insurance linked securities such as cat bonds – and to define them in a way that optimizes the use of capital based on program types, attachment points, and costs. Here the technology implication is the need for a reinsurance solution that allows the insurer to assemble a number of contracts and do the what if modeling to optimize the overall program. Reinsurance systems are increasingly needed to help model and craft reinsurance programs to optimize capital allocations and accurately deliver bordereaux reports – and more importantly, collect the reinsurance recoverables completely.

• Claims

– When a catastrophe hits, it’s all hands on deck. Assuring the claims system is intuitive so staff training and augmentation is fast becomes critical to address peak seasons. But also critical is making sure the insurer has consistent processes in place to address increasing regulatory scrutiny during and post catastrophe. And of course, like it or not, fraud often increases post catastrophe and so insurers look to expand their fraud capabilities. Claims systems need to have intuitive user interfaces to support rapid on-boarding of non-claims personnel during a cat situation. The ability to automate tasks helps to assure consistency in claims processing – which alleviates the regulatory pressure. And fraud tools, especially those that can access live weather data and geocode the losses against that data can help to identify potential fraud.

As you can see, there is a big need for technology to support these kinds of interventions. Key of course is data. Data is necessary to model the catastrophes and determine the impact on the book. While large insurers use one or more models across their different perils and geographies, smaller insurers tend to rely on their intermediaries/reinsurers to model for them. Either way, the need to access data is growing in importance. The more data available, the better pricing reinsurers provide. But beyond the modeling, insurers need underwriting and policy admin systems that can operationalize the models. E.g., if the model says obtain reinsurance for any frame building in protection class 7 of $650K or higher, having a system that fires off a task to obtain reinsurance will assure the insurer actually follows the model. And these systems need to be highly configurable so that as models change, underwriting can change as well.

We’re still early in the year. Catastrophe season won’t be in full swing for a few months and so insurers have time to continue to put their own plans together. Any catastrophe plan has to include both an element of capital management and technology.

Just as insurers are planning now for their own catastrophe season, I’m planning for a dry summer in my garden. I’m holding off planting new roses – although I have room for three more and they’re my favorites. I’m looking into grey water collection devices so I can reuse water from the shower or laundry. I’m already working to retune my drip system so that the most important plants get the minimum necessary to make it through the season. And key words – mulch, mulch, mulch. While it makes me sad to know that I’m bound to lose some of my garden,for now, today, I’ll enjoy the blooms.