Haiti Quake Prompts Jamaica Risk Model

An earthquake simulation model for Jamaica.

An earthquake simulation model for Jamaica.

Mona GeoInformatics Institute and the Earthquake Unit at the University of the West Indies (UWI) have developed an earthquake risk model for NEM Insurance Company Jamaica Limited, says NEM’s General Manager Chris Hind.

The model aims to determine whether NEM has adequate reinsurance coverage to cope with a 7.0 magnitude earthquake, similar to the one which struck Haiti in January 2010. Hind said such models are valuable to general insurance companies to ensure they have adequate coverage in the event of a disaster.

“Most catastrophe claims are actually paid from the capital of international reinsurers, not local companies,” the general manager told insurance brokers at the Terra Nova Hotel in Kingston on March 22. He said, “reinsurance coverage is finite and subject to event limits, which means we buy reinsurance up to a given level of loss. When that is exhausted, only our own capital and reserves remains to pay claims.”

Deciding how much reinsurance to buy and where to set event limits is thus crucial for a general insurer to operate successfully, he pointed out. However, such decisions have traditionally been guided by the output of international catastrophe models.

“Unfortunately whilst these models are sophisticated and widely accepted, they have limitations when it comes to accurately predicting the impact of catastrophes on small Caribbean nations such as Jamaica,” Hind outlined. The companies that develop the models have sketchy knowledge of local conditions and have not spent the time and resources to develop a detailed understanding.

“For them, Jamaica is not a lucrative market for catastrophe modeling,” he stated. Therefore,  locations, such as the South Eastern or Western United States get their attention, because there is much more value at risk.

“To close the gap, strengthen our risk management and protect our partners and customers, we have entered an alliance with Dr. Parris Lyew-Ayee and his team at Mona GeoInformatics,” he said. “This has resulted in several special studies to test the adequacy of our reinsurance.”

In December, Dr.  Lyew Ayee, Director of Mona GeoInformatics, worked in concert with Dr. Lyndon Brown, head of the Earthquake Unit at UWI, to create a model that simulated the Haiti Earthquake along the Enriquillo-Plantain Garden Fault at the worst possible position for Jamaica. The focal point of the model earthquake was set at Ness Castle-Arntully, in the foothills of the Blue Mountain range.

NEM and UWI team.

Dr. Parris Lyew Ayee (centre), director of Mona GeoInformatics, discusses Jamaica’s first indigenous catastrophe simulation model with Chris Hind (right), general manager of NEM Insurance Company Jamaica Limited, and NEM’s deputy chairman, Kathy Moss.

“We created a Haiti-in-Jamaica scenario, with the same depth and magnitude,” Lyew-Ayee stated, “and building types were classified by roof and wall types, with weights applied to their sum insured value, as well as the probability for loss in category and each zone.”

“The simulation revealed that the entire island would experience some level of shaking, with more than 75 per cent serious enough to cause structural damage in some buildings,” he said, “and a significant number of people and buildings are located within this damage zone, as well as a significant number of NEM clients and sum insured values.”

The model demonstrated that the company’s reinsurance provisions with overseas providers would adequately cope with such a disaster, as well as a second catastrophe in the same year, Hind said.

“We, therefore, plan to examine our property portfolio, and educate our clients about the probabilities that this simulation reveals; and, determine, based on the geology of their locations and the structure of their buildings, how best we can refine our levels of exposure.”

     

Mark Lee

About Mark Lee

Editor, author and writer with career spanning print, radio, television and new media.

%d bloggers like this: