2017 and 2018 were difficult years for the reinsurance industry. 2017 saw losses for property insurers and reinsurers reaching $150 billion (the highest ever) as a result of three major hurricanes making landfall in the U.S., and 2018 saw insured losses of roughly $80 billion, mainly caused by California wildfires and Typhoon Jebi in Japan. The losses increased concerns investors had about the impact of climate change on the industry.
My column of Sept. 21, 2018 discussed several key issues including some of the historical evidence, recency bias, and avoiding the mistake of confusing information with value relevant information.
Given the recent losses and heightened concerns, I thought it important to take a deeper dive into the science of climate change as it relates to reinsurance risks.
Top Climate Change Risks Don’t Impact Reinsurers
In October 2013, three billionaires, Michael Bloomberg, Hank Paulson and Tom Steyer, commissioned a study called Risky Business. They hired a leading economic consultancy, the Rhodium Group, who partnered with Risk Management Solutions, the leading global catastrophe modeling firm, to conduct an objective study of the threats of climate change to the U.S. economy.
In their report “The Economic Risks of Climate Change in the United States,” they identified the four largest economic risks to the U.S. As you will see, three of the four issues do not impact the reinsurers.
- Threats to coastal property and infrastructure. Over the next 15 years, the average annual cost of coastal storms along the Eastern Seaboard and the Gulf of Mexico could increase by $2.0 billion to $3.5 billion. Potential increases in hurricane activity could increase the average annual cost by another $4.0 billion. While this scenario would clearly be bad for the economy, it seems likely it would have limited impact on the reinsurance industry for several reasons. First, roughly half the increase is driven by flood risk associated with rising sea levels. Residential flooding is currently reinsured primarily by the federal government through the National Flood Insurance Program, which buys only a limited amount of reinsurance (most losses are borne by the U.S. taxpayer). While commercial flooding is covered by reinsurers, coastal flooding due to sea levels rising is a creeping risk rather than a sudden risk. According to the Risky Business report, if we continue on our current trajectory, sea levels in New York City would rise 2.4 feet to 4.2 feet by the year 2100. This would mean an average of 0.95 inches per year. Because reinsurance contracts renew every year, reinsurers can see this increase in risk coming, and can react in time to incorporate it into pricing and underwriting. Finally, the increase in potential loss from increased hurricane risk is small relative to the existing hurricane risk.
- Rising temperatures reduce labor productivity. By 2050, the average American may experience 27 to 50 days above 95 degrees Fahrenheit (a two to three times increase). This means that in some regions it may be too hot for people to work outside during the hottest part of the day. This would impact labor productivity in such industries as construction, agriculture and utility maintenance. While this scenario could have a material impact on affected industries, these economic risks are not insured or reinsured.
- Strains to our energy system. As the country experiences hotter summers, demand for air conditioning will increase as well, putting a strain on regional generation and transmission capacity, and perhaps drive up energy prices. Again, none of these additional costs are covered by reinsurers.
- Health impacts from heat. The worst health impacts from increasing heat will be felt by the poor, who cannot afford air conditioning, and the elderly, who are too frail to withstand the increased heat. The combined impact of heat and humidity can be dangerous or even deadly. This scenario poses economic and social risks, but the burden will fall on the health care system, rather than on property insurers or reinsurers.
Let’s now turn to another important issue: Headlines are often misleading because the media’s goal is to draw attention/interest—if it bleeds, it leads.
Let’s look at a Jan. 10, 2019, article from the New York Times. The headline trumpets “Ocean Warming Is Accelerating Faster Than Thought.” They present a graph showing that the shallower (between 0 and 800 meters) ocean heat content has been increasing dramatically over the past 80 years. Since warmer sea surface temperatures provide more energy to hurricanes, you’re left to draw the conclusion that hurricanes are also increasing in frequency or severity.
However, oceans are warming more rapidly near the poles than near the equator. In terms of U.S. hurricane risk to reinsurers, what we care about is the sea surface temperatures in the tropical regions of the North Atlantic. Data from the National Oceanic and Atmospheric Association (NOAA) website shows there has been no trend at all in North Atlantic sea surface temperatures. In fact, they are basically the same as they were from 1940 through 1960, and they oscillate in a cycle called the Atlantic Multidecadal Oscillation (AMO). According to NOAA, the cause of AMO is a decades-long cycle of ocean waters circulating from north to south and back again.
Using the data, Stone Ridge has found that, while there was a 0.46 correlation between rising sea surface temperatures and hurricanes forming, there was very little correlation (just 0.16) between rising sea surface temperatures and landfalling hurricanes.
The explanation for this lack of correlation is provided by the scientific theory suggesting that wind shear increases when sea surface temperatures are warmer. Wind shear happens when winds in the upper atmosphere are moving at a different speed from winds at a lower altitude. This effect tends to break up hurricanes before they can make landfall.
Adapting To Wildfire Risk
There’s another important point we need to cover: Insurers adapt to increasing risks by changing underwriting requirements.
Wildfire is one risk where climate change may be impacting the severity of insured losses. Prior to 2017, the costliest wildfire years in the U.S. caused less than $2 billion of insured losses.
However, 2017 and 2018 saw insured losses from wildfires of $15 billion and $19 billion, respectively—wildfires have become not only more frequent but also costlier over time. Some of this is due to forestry practices, and some may be linked to climate change. In addition, we continue to build homes in the wildland-urban interface, the areas with the highest exposure to wildfire risk.
The insurance industry has reacted in several ways. First, it has now incorporated the greater risks into risk models, justifying significantly greater premiums. Second, it has adapted underwriting standards.
For example, one major insurer has started refusing to renew policies in high-risk areas unless homeowners conform to California Department of Forestry and Fire Protection (CAL FIRE) guidelines for wildfire safety. These guidelines require that a 10-foot radius around the home, the land be kept “lean, green, and clean,” cleared of anything flammable, such as trees and bushes.
Through stricter underwriting requirements, primary insurers are reducing the risk of insured losses to wildfires.
Before summarizing, we need to cover one more important issue that might prevent you from investing if you are unaware of it.
“Recency bias” is the tendency to overweight recent events/trends and ignore long-term evidence. This leads investors to buy after periods of strong performance—when valuations are higher and expected returns are now lower—and sell after periods of poor performance—when valuations are lower and expected returns are now higher. This results in the opposite of what a disciplined investor should be doing: rebalancing to maintain his or her portfolio’s targeted asset allocation.
The aforementioned three major U.S. hurricane landings in 2017 resulted in significant losses for the reinsurance industry. SRRIX lost 11.4% in 2017. The fund also lost 6.1% in 2018.
With that in mind, let’s look at the historical evidence, recalling that global warming is a very long-term phenomenon, one that doesn’t occur overnight but rather is felt over many decades. In the three years prior to 2017, the fund returned 11%, 7.9% and 6.4%, respectively. It’s also important to note that, before 2017, no Category 3 or greater hurricane had landed in the U.S. since 2005. More importantly, there is no evidence of any long-term trend in hurricane landfalls.
When it comes to matters affecting reinsurance, the measure of hurricane risk we should focus on is the frequency of landfalling hurricanes. The good news is that there is a clean dataset (HURDAT, maintained by NOAA) on landfalling hurricanes going back to 1851.
Meteorologists and climate scientists have studied this dataset intensively and have found no significant trend in the frequency of landfalling hurricanes over time. Data shows the long-term average is about 1.8 U.S. hurricane landfalls per year and about 0.6 major (Category 3 or greater) hurricane landfalls. The evidence also shows that one year’s losses from hurricanes tell us nothing about the following year’s losses.
The past two years were the worst two consecutive years in history for insured losses. It has also been an opportunity for the mainstream media to draw connections between climate change, natural disasters and insured losses.
While some investors exit the asset class because of fears of climate change, their doing so creates an opportunity for the evidence-based investor to dive deep into the data and facts to get a clear picture of the actual impact of climate change on the reinsurance asset class. As discussed above, the actual impact of climate change on property reinsurance is less than the popular perception.
The bottom line is that the logic of investing in reinsurance as an asset class still holds, whether or not you believe in the effects of global warming. Also note that the industry historically has had its strongest performance in years immediately after large losses. The reason is simple.
Losses create a need to restore capital, and the industry raises prices to address that need. This year is no different, as premiums for loss-impacted reinsurance contracts have risen between 5% and 35% depending on location.
On a final note, because losses typically are greatest during hurricane season (June through November), most premiums are allocated to that period, meaning a reinsurance fund tends to produce the highest returns in the second half of the year. Of course, that is also when losses are likely to be greatest. (Full disclosure: my firm, Buckingham Strategic Wealth, recommends Stone Ridge funds in constructing client portfolios.)
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 130 independent registered investment advisors throughout the country.