Earlier this week we wrote that the Caribbean Catastrophe Risk Insurance Facility (CCRIF), the catastrophe risk pooling facility who provide parametric based disaster insurance cover to Caribbean countries, had finalised their renewals for the 2012-2013 policy year. All 16 of their member countries renewed their hurricane and earthquake cover and now the CCRIF are actively negotiating and discussing their soon to launch excess rainfall cover with their members.
The excess rainfall product has been long-awaited as it will fill a coverage gap for tropical storms which strike Caribbean islands but do not reach the wind speeds necessary to trigger the parametric hurricane policies. Rainfall and the resulting flooding and landslides from tropical storms which do not attain hurricane status can result in significant losses for Caribbean countries so it’s expected that the excess rainfall policy will be well received.
The excess rainfall product will provide index-based cover to those who take up the policies. Reinsurer Swiss Re helped to develop the parametric rainfall index for the product using available NASA-processed satellite data. Eventually the product hopes to be integrated with a numerical rainfall model making the index and parametric measurement of claims even more accurate.
Detailed information on the model was published previously by the CCRIF which we reproduce here for your interest. These are quotes from Simon Young, CEO of Caribbean Risk Managers Ltd. (CaribRM) the facility supervisors for the CCRIF, taken from a interview for the CCRIF’s newsletter in April.
The CCRIF/Swiss Re XSR model uses NASA/JAXA Tropical Rainfall Measurement Mission (TRMM) daily rain data to compile a 5-day running aggregate of rainfall measurements at all of the TRMM grid nodes across a given country. As used in other CCRIF products, the Multi-Peril Risk Estimation System (MPRES) exposure database is utilised to map exposures across a country at 30 arcseconds (~1 km) resolution. Remote sensing data, economic and demographic statistics for 2010 are used to generate the exposure database. The database is designed to provide acceptable estimates for losses to physical assets from hydro-meteorological and geophysical hazards. Since the TRMM nodes are at ~25 km resolution, the 1 km MPRES exposure data is mapped onto the TRMM grid. The value of each exposure node is distributed between the closest four TRMM grid nodes and weighted by distance. This provides a distribution of the total MPRES values between the rainfall measurement points covering each country. For scaling purposes, 1% of the total MPRES exposure value is used as the base XSR exposure.
To calculate index losses for both historical and real-time analyses, a 5-day aggregate rainfall is calculated for each TRMM grid node using a moving window, which ensures that peak measurements are captured. A rainfall event occurs when the 5-day aggregate exceeds 50 mm and ends on the day before rainfall next falls below 50 mm. Events are logged for each TRMM measuring point. For each event at each TRMM grid node, the single highest 5-day aggregate rainfall measurement is used to calculate the index loss rate via a vulnerability curve which maps indemnity percentage to rainfall amounts.
The indemnity rate for each event is applied to the exposure value of the TRMM grid node, to give the individual index loss for the event for the grid node. To calculate the national index loss, the individual index losses at each grid node are added together each day. National-level events are defined as continuous periods where there is an ongoing event at one or more TRMM grid nodes. Therefore, once an event occurs at one or more of the TRMM grid nodes, a national loss is assigned to it with the date of the last day of the event as the event identifier. National losses are also aggregated on an annual basis, thus allowing coverage to be offered on a per-event or on an annual aggregate basis at the national level.
For each CCRIF country, the steps described above are taken to calculate index losses. Application of historical rainfall data to the exposure database using the vulnerability function generates a historical risk profile (event-specific and annual aggregate). This risk profile provides the necessary information for CCRIF/Swiss Re to price coverage. Coverage characteristics, within limits, are selected by each country separately (in the same way as existing CCRIF earthquake and hurricane coverage selections are made). Premium cost is risk-based, so depends on the rainfall risk profile of the country and the coverage characteristics selected.
The extreme rainfall product is triggered independently of the current hurricane product, and if both policies trigger then both payouts are due. The current hurricane policy is linked to wind and storm surge damage in a defined Tropical Cyclone. While the excess rainfall product can be triggered for a Tropical Cyclone, it can also be triggered in non-cyclonic systems if the rainfall trigger thresholds are met.
Rainfall risk profiles are produced by CCRIF as soon as the background exposure and vulnerability data are generated by Swiss Re. This is being done on an ongoing basis, and all risk profiles should be available within the coming months. Risk profiles are being produced and coverage will be offered to Guyana and Suriname as well as to other Caribbean countries not currently members of CCRIF. Once rainfall risk profiles have been developed, CCRIF will discuss coverage options with each country individually and policies will incept once coverage levels have been agreed.
We asked Simon Young what the latest status with the rollout of the excess rainfall policies was. He told us “We are in the relatively advanced stages of negotiating coverage with Jamaica and are also in the preliminary stages of discussions with Haiti and Trinidad & Tobago, and hope to be commencing discussions with other CCRIF clients over the coming few weeks.”
That sounds encouraging and it seems likely that the 16 CCRIF members will all have a chance to bind an excess rainfall policy before we get too far into the 2012 tropical storm season.