- What can I do to estimate losses for the current policy year?
- Why can’t I estimate losses for the current policy year after three months of loss development?
Loss Forecaster relies upon benchmark loss development factors (LDFs) to develop current incurred and paid losses to their estimated ultimate value. Loss Forecaster chooses what LDF to apply to each policy year based upon the coverage and age of the period. The age of an annual loss policy is the difference between the current evaluation date and the policy inception date, usually measured in months and defined as “Months of Loss Development.” Most benchmark LDFs are published on an annual basis beginning at 12 months of development. In order to allow mid-year analyses that do not rely strictly upon year-end evaluations, SIGMA actuaries have interpolated LDFs for 12 through 180 months of development for use within Loss Forecaster. In addition, Loss Forecaster allows the user to apply LDFs to a policy period with as few as 6 months of development.
It should be noted that SIGMA actuaries recommend against relying solely upon developed losses for any policy period with less than 12 months of loss development. Instead, consideration should be given towards projected losses, which are typically based on long-term averages of loss experience in historical policy years. An example of how to use this suggested methodology is covered in the following Lunch & Learn video: Lunch & Learn Webinar: Completing a Mid-Year Analysis in Loss Forecaster
As to the final question, even the LDFs from 6 to 12 months can cause unstable results. Anything less than 6 months would be increasingly erratic. 6-month LDFs for paid losses typically range from 10.000 – 25.000 depending upon the coverage. In practice, this means that a single $1,000 paid loss can lead to an additional $25,000 of estimated reserves. Stated another way, an LDF of 10.000 implies that 10% (1/10.000) of the actual development within a policy year has happened thus far. So, using an LDF of 10.000 is the equivalent of using information from 10% of a policy year to predict the behavior of the other 90%. This is known in actuarial terms as a highly-leveraged LDF, and its use is frowned upon. Instead, the loss projection methodology suggested above, which is based upon experience from a much more stable period of time, is recommended.