The 2021 Natural Catastrophe Bingo Game. What is in the 2022 roller?

By: Andrew Siffert

Natural catastrophes are certainly not a game; but, like any bingo game, each draw of the ball from the roller provides a surprise to fill our card. There is no doubt that 2021 had a few surprise natural catastrophes for the insurance industry, including a late draw on December 30th of a front-range billion-dollar wildfire. This BMS Re insight is not to highlight all the surprises that occurred, but to take a look at what is left in the roller for 2022.

There is no way to predict natural catastrophes, but looking back at the events of 2021 there were surprise gray swan events like winter storm Uri and the following massive power failure that resulted in the biggest winter storm loss ever for the insurance industry. This purely man-made event clearly got 2021 off to a higher-than-expected loss year, when the first quarter is typically quiet for the insurance industry. The winter storm losses accounted for 20% percent of the U.S. insured loss, which is highly unusual. There is no doubt that if the electricity would have stayed on, the losses would have been just a fraction of the overall loss that was experienced as winter storms typically only accounts for 6% of the U.S. average annual loss. However, the overall failure of the power grid is just another example of how critical infrastructure is increasing insured losses and compounding event losses. This trend seems to be growing as infrastructure ages and becomes more strained due to demand. So, in 2022, we should not be too surprised if other compounding events like this occur again due to the failure of the existing infrastructure. There is a clear need for the insurance industry to contemplate what risk adjustments need to be made to account for these types of compounding events, which are not part explicitly of a catastrophe model output.

2022 Hurricane Bingo Balls

Hurricane Ida is a good example of compounding factors. The loss of a major T&D line into the city of New Orleans prolonged the power outages for many risks within the city of New Orleans compounding business interruption losses and the recovery efforts. Ida’s heavy rains into the Northeast also showed how over development and aging infrastructure, in some cases, cannot keep up with flooding rains which appear to be becoming more frequent. It was clear going into the 2021 Atlantic Hurricane season that an active year was to be expected and the basin was active with 21 named storms. Having said that, our BMS Re Insight that summarized the season mentioned that most of these storms were short-lived and weak. Active storm years also correlate to higher chances of named storm landfall, which is what matters to the insurance industry. Given 8 of the 21 named storms made landfall in 2021, it was an extremely active year for named storm impacts. However, with only two hurricane landfalls, this is in line with what should be expected since the long-term rate is about 1.6 hurricanes per year. Perhaps what is more interesting is that 2021 came close to ending the hurricane drought that is occurring in New England. Henri came within 41 miles of ending the New England’s hurricane landfall drought that started with Hurricane Bob (1991). What might be a bit scarier is it has been 83 years since the last major hurricane landfall (Long Island Express (1938)), into the Northeast. To compare, the expected recurrence interval is every 30 – 50 years, based on different methods of calculating return periods based on the long-term record and paleoclimate sediment records. The main concern here for the insurance industry is that many areas in the Northeast have experienced significant exposure growth and have not had their defenses tested in decades. Remnants of Ida and tropical storm Henri, which accumulated over $6 Billion in insured loss to the Northeast, provides a learning opportunity on how the region will fare when the next hurricane occurs to the region.

BMS iVision and the Verisk Respond product are great resources for the insurance industry in helping to understand wind impacts. This is an overall aggregated summary of all the 2021 wind swath from the landfilling named storms this season. Eight named storms impacted the U.S. this season. Two of these named storms were hurricanes at landfall (Ida and Nicholas). Ida was a very powerful hurricane with wind gusts topping out the scale along the Louisiana coastline, which is indicated in the wind swath map. Henri provided some strong winds into New England and for the third season in a row, the Florida peninsula missed most of the action this year.

It is not just the New England or the Northeast that is awaiting the surprise of a hurricane in 2022, the entire East Coast has not seen a major hurricane landfall since Hurricane Jeanne in 2004. So will the 2022 bingo roller draw such an event? It is too early to say, but all eyes are on El Nino Southern Oscillation (ENSO) at this time, which suggests the strong La Nina currently occurring will weaken into ENSO neutral conditions for the August – October time frame (45% chance). Remember an El Nino typically reduces Atlantic Hurricane activity by increasing wind shear and a reduction in named storms tends to mean less named storm impacts. This is not called for at this time. Looking solely at ENSO and U.S. mainland major hurricane strikes the cool neutral phase of ENSO has been by far the most favorable for these landfalls which could be the ENSO phase during the 2022 Atlantic season. So, just based on the projected ENSO conditions, there could be more hurricane landfalls which continues to correct for the lack of landfall between the years of 2006 – and 2016 and just maybe the East Coast might get into the landfall action.

What Severe Thunderstorm Surprises Continue in 2022

Many in the Insurance industry might have been surprised by the active, costly, December 2021 severe weather outbreaks, which accounted for 20% of the overall U.S. thunderstorm loss in 2021. This costly year might be interesting because of the frequency of events over $25 million of insured loss which has been rising exponentially since 2010. 2021 had 57 such events which is 183% above the expected average of 31 events between 2010 and 2020. This is unusual because according to local storm reports from NOAA for tornadoes and high winds, damage reports were just below average.

This is the PCS BMS view of hail frequency that occurred in 2021 across the U.S. THis is not all hail events, just hail events that cover the Verisk PCS assigned dates.

The hail peril tends to drive the average annual loss for the insurance industry, but they were at the lowest reported level since 2005. However, this year the industry experienced a surprise event in what was a relatively quiet night severe weather-wise. On April 28th, 2021, three separate supercell thunderstorms popped up and hit San Antonio, Dallas, and Norman resulting in a costly evening for the insurance industry.

This is the NOAA Local Storm Report (LSR) departure from the average number of reports that are expected at the National Weather Service offices across the country. The blue colors represent less than normal severe weather reports. Yellow represents more than normal severe weather reports. This year there was more than normal severe weather in the Northeast U.S. which could explain the higher observed loss given the increase in exposure and the increase in the frequency of events in this area.

Overall, the Central Plains had a relatively quiet year in terms of severe weather with parts of the Mid Atlantic seeing more observations of severe weather than parts of Central Kansas, which might be a surprise to some. This could also explain the higher frequency and insurance-related events and the higher loss given the exposure is much higher along the East Coast than it is in the Central Plains. The tornadoes that occurred in December were unprecedented in the historical record. Although the current official rating of the Mayfield, KY tornado stands at an EF4 this means the bingo space of an occurrence of an EF5 tornado is still open and waiting to be drawn as the long-standing EF5 tornado drought for the U.S. continues. This is currently the longest period between occurrences in the recorded catalog. However, this does not mean a large EF-5 tornado cannot hit a community. It is just a matter of time before it happens, which is why it is important to understand tornado accumulation which is what we do within Pathlight Analytics using BMS HiRES.

Wildfire Bingo Spaces Left Open

The occurrence of large destructive wildfires in the Western U.S. seems to be almost like a free spot on the natural catastrophe bingo card, but the list of other spots that could be filled is long. The Marshall Fire that impacted Louisville and Superior, CO on December 30th is just the latest example of large wildfire losses that can occur outside of mountainous terrain of the western U.S or outside of what would be the normal wildfire season. It seems every few years now an event like the Bastrop, Gatlinburg, or Fort McMurray wildfire can occur and cause loss to the insurance industry. But there are other tinderboxes just waiting given the right conditions. Communities along the Pine Barrens of New Jersey, the encroachment of the suburbs of Maimi to the Everglades, or the thousands of towns in forested areas in the northern tier states are just a few examples of risks that could cause insured losses given the right fire conditions like witnessed along with the front range of the Rockies this December. However, there is a growing list of data analytic companies that are helping the insurance industry understand this hazard. In fact, BMS Pathlight Analytics has a dialog with 13 such vendors; more than any other natural catastrophe peril.

Market Place of Ideas

Natural catastrophes are full of surprises, but they are often familiar. Recently, Roger Pielke Jr. wrote in his blog that we focus too much on familiar risks. Our catastrophe model helps understand these familiar risks that the insurance industry has come to expect based on experience and knowledge from past events. We have seen the catastrophe risk models fail at emergent risks such as the winter storm losses of February 2021. It is these surprise compounding events that appear to be the new focus to the insurance industry and catastrophe model vendors. Nonetheless, the Hunga Tonga volcanic eruption on January 15th, 2022 showed that extraordinary events also deserve some attention. This past year an insight was written about how a geomagnetic storm is one of these extraordinary risks that the insurance industry needs to better understand. The analogy of natural catastrophes as a bingo card is similar to what Roger describes as the “Marketplace of ideas”. As mentioned above, we may not be able to predict when future natural disasters will occur, but as an industry that is meant to provide the financial stability to recover from natural disasters, it is critical to continue pushing forward with understanding what can happen in the future. The 2022 bingo card has many different surprises disasters just waiting to happen. Disasters are a serious problem and with every disaster, life can be impacted directly or indirectly. Some of these surprises in 2022 will be familiar, some are emergent risks with compounding impacts and others are extraordinary events that are almost unthinkable. The use of catastrophe risk models can help understand some aspects of the unthinkable. When the models don’t have a solution there are pragmatic ways to understand unthinkable impacts. However, the free space on the 2022 bingo card is taking steps to reduce societal vulnerabilities to disasters is worth pursuing and is picking up momentum. This alone will help no matter what happens with future disasters.