By Andrew Siffert – Vice President / Senior Meteorologist
There is no question that 2020 was a unique year. Many could not wait for the year to end and numerous summaries highlighting the many different catastrophes and their impact on the insurance industry have been published. However, there is a need to put things into perspective at times and this BMS Insight looks to highlight a few unique aspects of 2020 and the catastrophes that have not been widely reported that have positively impacted the insurance industry.
There is no escaping all the news about all the catastrophes in 2020 that have totalled $83B according to Swiss Re and $82B according to Munich Re. However, what is not widely mentioned in all of the various discussions is how normal these losses are. According to AIR Worldwide (AIR) and its 2020 Global Modeled Catastrophe Losses report the average annual loss (AAL) is modeled at $99.6B. The 1 percent aggregate exceedance probability insured loss (or the 100-year return period loss) is nearly $301B. Now with that being mentioned, this is not a complete apple to apple analysis as there are catastrophic events that occurred in 2020 that were not modelable events; however, the broader point is overall losses appear to be what the insurance industry should expect in any given year.
It is also a good reminder not to put much weight on trends in economic and/or insured losses and to measure catastrophic trends in the perils that cause the loss because there are just so many moving parts. Roger Pielke Jr., a professor at the University of Colorado, continues to lead this area of analysis by putting these losses into perspective. Along with the comments from the Intergovernmental Panel on Climate Change, which mentions “Increasing exposure of people and economic assets has been the major cause of long-term increases in economic losses from weather- and climate-related disasters (high confidence)", Roger Pielke Jr. for over a decade has analyzed Global disasters as a proportion of global GDP data provided by Munich Re and the WorldBank. The latest research continues to headline a long-term downward trend and even with Global GDP declining due to COVID-19 it shows that there are loss complexities due to social-economic issues, but overall they likely could have been worse. The following three sections are just three highlights as to some of the reasons which have not been mentioned in a lot of these annual reports.
Severe Weather-Related Observations
In reading the various reports most of the largest and costliest natural disasters in 2020 occurred in the United States. Severe Weather in the U.S. leads to over $30B in insured loss across 40 separate industry-wide events according to Property Claims Services (PCS). However, what is interesting is when you look at the severe weather data from 2020 it was not a record year in terms of reports of severe weather. Therefore, there is a disconnect between the observations of the peril of severe weather and the large amount of insured loss that occurred. For example, observation of hail occurrences where historically low since 2005, and considering an estimated 60 percent of average annual storm losses in the U.S. are a result from hail, it would follow that most losses could have been driven from non-tornadic wind activity.
The Nashville tornado on March 2 once again proved that cities and large expansive metropolitan areas are not shielded from large tornado impacts. However, like hail, total tornado counts were observed just below the annual mean count. The U.S. currently is in the second-longest period between observations of EF5 tornadoes. The last EF5 tornado observed in the U.S. was on May 20th, 2013. The other unusual aspect of the severe weather season was Florida and Illinois each saw more tornadoes than Kansas and Oklahoma combined in 2020. And, since tornadoes are correlated with outbreaks of other severe weather perils, maybe it was severe weather in the more populated southeast which also contributed to the higher amounts in insurance loss this past year.
Not all severe thunderstorm wind impacts are a result of tornadoes, the Midwest derecho was by far the most impactful event of 2020, which has been dubbed the equivalent to a major hurricane impacting the state of Iowa. Based on reports of high winds and wind damage, 2020 was on par with 2019 and overall the third-highest year of wind local storm reports since 2005. Therefore, maybe a higher percentage of insured losses are coming from severe weather wind events.
Major Earthquake Drought
It probably is already in hindsight, but 2020 started with the most powerful earthquake to hit Puerto Rico since 1918, when a large Mw 6.4 earthquake struck just 5 miles off the southern coast on January 7th. Another rare earthquake struck about 9 miles west of Salt Lake City as an Mw 5.7. Although the highest earthquake hazard in the continental United States is largely associated with California, which still awaits the big one, earthquakes in the past have had devastating impacts. However, the 2020 global earthquake catalog was quite kind to the insurance industry with the second year in a row with a relatively low number of Mw 7.0 or greater earthquakes from around the world.
Named Storm Activity
Many of the 2020 reviews have highlighted how active and devastating the 2020 Atlantic hurricane season was in particular on the central Gulf Coast. This active landfall season had an impact on the total insured losses in the U.S., which currently accounts for just under $20B and should continue to develop as claims are still being processed. In fact, in the season-ending tropical update, it was noted that the clustering of storms was very evident this year. However, the insurance industry needs to understand that the hurricane magnet Florida was relatively unscathed during the active season and this is the 28th year that the very populated southeast Florida coastal city of Miami has not been hit by a major hurricane directly. Although Isaias had its impact widely felt in New England as it transitioned into an extratropical storm, this part of the U.S has not had any direct impacts from a hurricane since Bob in 1991. We have seen several hurricanes impact Nova Scotia so it is certainly possible. Lastly, it has been 31 years since the last major hurricane has made landfall in South Carolina, which was hurricane Hugo, 1989. Remember the U.S. went through a period of 12 years without a major hurricane impact and the landfall activity we are seeing over the last three years now may just be nature's way of reverting landfall statistics to a Poisson distribution. Unfortunately, some areas are long overdue for a major hurricane landfall.
Spring 2021 Early Look Ahead
2020 is now in the rear view mirror and half of the meteorological winter is also now behind us. Looking ahead, the early trends for spring show the presence of La Nina in the weather patterns. This means that we need to keep a close eye on the coming months, as La Nina is known to cause stronger severe weather seasons in the United States.
La Nina is a cold phase of a large oscillation in the central Pacific Ocean, called ENSO, (El Nino Southern Oscillation), which alternates between cold and warm phases. Preliminary data shows that the ENSO has already reached the peak of its cold phase. While it was not the strongest of all time, this La Nina was indeed quite strong, which increases the likelihood of its impact reaching into spring 2021.
One of the main influences of La Nina (or any other ENSO phase) can be seen in the changing jet stream. The image below shows the average position of the jet stream during the La Nina seasons and the corresponding weather development over North America. The twisted jet stream brings colder air and storms down from Canada into northern and the north-western United States, and warmer and drier weather to the southern parts.
The image below shows a comparison of hailstorm and tornado frequency during the spring season in the United States, between El Nino and La Nina years. It is quite interesting to see that in a La Nina spring season, there is a substantially higher frequency of hailstorms and especially tornadoes in the “Tornado Alley” south-central plains of the United States.
The cause of this has a lot to do with the weather pattern. As La Nina promotes a high-pressure system in the North Pacific, which usually corresponds to the pressure drop over western Canada and the north-western United States. This can act as a source for frequent cold fronts that move from western Canada down towards the south-central United States, where they meet warm moist air coming from the Gulf of Mexico. Together with the altered jet stream over the United States by the La Nina, this produces a combination of conditions needed for the storms to become severe and tornadic.
Below is the season forecast from the European Centre for Medium-Range Weather Forecasting (ECMWF) model. The forecast suggests the pressure forecast will be lower in western Canada, as well as in southern and southwestern United States. This low-pressure zone means a stronger warm and moist southerly flow from the Gulf of Mexico into the south-central United States, as a low-pressure system spins counterclockwise.
Below is a graph which shows annual tornado (EF1+) numbers in the United States from 1950 to 2020. Looking at the years, we can see that many of the most active tornado years were La Nina years. Therefore, if this spring La Nina holds together, given that it is the most active part of the year for tornadoes, and the occurrence of tornadoes is correlated with other forms of severe weather there is a good chance the central plains could get back to seeing severe weather which was lacking last year. This is clearly something the BMS Analytics team will be watching and helping clients with tools like BMS IVision to help understand severe weather impacts.