We provided some early opinion following the landfall of Hurricane Ida onto the US mainland in Louisiana on 29 August 2021. Following some conversations with underwriters and market personnel, we are able now to provide a little more context. 


Hurricane strength winds quickly dissipate as they cross over land, as the storm loses its power source of warm sea water. Hurricane Ida became a post-tropical cyclone as it moved North. The sustained winds of 175mph (275km/h), which caused devastation in Louisiana subsided  to gusts of around 40mph. These more moderate, and therefore less damaging, winds, do nonetheless spin off tornadoes, which can wreak havoc on a localised basis. However, more severe was the heavy rainfall across much of the north eastern USA as Ida continued its route. In Central Park, New York, more than 3 inches (76mm) of rain fell in less than one hour on 1 September (this broke a record less than one month old, set when Tropical Storm Henri passed over the city, dropping 2 inches of rain). 


This heavy rain has caused severe flash flooding adding to the overall cost to the insurance industry. The nature of losses for a flood and wind-driven water event is different to the profile of wind damage losses that took place in Louisiana. The greater population density in the North East increases the number of vehicles subject to the loss and automobile damage will be a factor in the overall cost to insurers. 


Flood is written by a number of Lloyd’s syndicates, both for commercial policyholders (on one of an all risks policy, on a flood only policy and sometimes on a difference in conditions policy picking up perils specifically excluded under a fire policy). Homeowners generally buy flood insurance from a US government entity called the National Flood Insurance Plan (NFIP) although some commercial insurers including Lloyd’s syndicates are able to offer better terms than the federal programme. The NFIP also buys some commercial reinsurance (including Lloyd’s syndicates and insurance linked security writers) to limit its overall loss. It is difficult to gauge the overall level of flood losses at this stage, although underwriters feel the images to be more dramatic that the actual damage. Again, many homes were left without power and the death toll has climbed to around 50 people. 


Catastrophe modelling firms, some insurers and reinsurers and some brokers continue to make projections of the overall cost to insurers from the storms. These currently sit in a range from around US$17 billion to more than $30 billion. The underwriters we have spoken to have cautioned that these early estimates need to be tested over time. Many recent storm losses have been subject to claims’ inflation and so-called demand surge (the rapid even if temporary increase in the cost of raw materials required for building repairs).


While the actual aggregate insured loss is only of academic interest, it does provide a useful shorthand for the industry. Individual insurers will chose a reinsurance programme to fit their underwriting strategy and risk profile. So smaller local, insurers will expose their reinsurance programme at smaller levels of market loss. Large regional and nationwide insurers will begin to incur reinsurance recoveries at market losses of $10 billion. Some reinsurances incorporate a market loss warranty, and will only respond when the total loss exceeds a pre-defined threshold.  


This flooding and rainfall in the North east event will inevitably increase overall claims to reinsurance programmes. Rather than be treated as a separate event, with a second net retention for primary insurers exposed in both the South and North East of the USA, claims look set to aggregate into one reinsurance recovery. The hours clause, the period in which claims from a single cause can be aggregated into a single reinsurance claim, has tended to increase over the past few years. Many reinsurances now contain hours clauses of 168 hours (7 days), with named storm cover a frequent alternative. The latter allows all claims with the same storm as proximate cause to be aggregated together for reinsurance recoveries.  


As regards Lloyd’s exposure, it is still really too early to tell. The first estimate of loss by syndicate will be included in the third quarter syndicate statistics pack, due in November. Our preferred syndicates have a very strong record in cautiously forecasting the impact of any event on their books, dating back from Harvey, Irma and Maria in 2017, to Katrina in 2005 and the 9/11 events of 2001. 


There have been material events so far this year, with the most significant aside from Hurricane Ida being the winter freeze in Texas in February, otherwise known as Storm Uri. Despite these two significant events, and other smaller losses such as the European Floods, political unrest in South Africa and flooding in  New South Wales, we expect syndicates to remain within catastrophe budgets (the estimated annual provision for the cost of large and catastrophe claims). However, there we are now at the peak of the Atlantic storm season, so additional events could change that view. In recent years, the catastrophe budget has been around 10% of net premium to the Lloyd’s market. 


Under Lloyd’s underwriting account conventions, claims are allocated to the underwriting year in which the policy incepts. A rule of thumb for hurricane claims is roughly 85% to the current underwriting year (in this case, 2021), and 15% to the previous year (2020). There is often a small proportion allocated to the oldest open year, with construction risks and some parts of the delegated authority book still on risk. 


Separately, Lloyd’s is reporting its half yearly results, to the end of June, on Thursday 9th September. We will report on these with our comment in due course, although those interested can register to see the market results briefing here