General liability rating

The general liability line of business varies from the standard rating algorithm because most general liability costs relate to what are referred to as exposures. Like the workers’ compensation line, many general liability classification exposures (also known as basis) appear in the user interface in window mode of the full policy term. To properly rate the policy, you may need to specify what portion of the full term exposure applies before and after a rating change. For example if 70% of a store's sales is in December, you want 70% of the sales exposure to use the rates effective during that month.

Some general liability premiums are determined with rate-scalable costs with effective dates like standard lines of business.

Thus, the rating engine for general liability is a hybrid between the workers’ compensation approach (all costs are unsliced) and standard rating approach (mostly by slice dates).

The structure of general liability data also ensures that the general liability rating engine cannot rate merely from a slice date forward.

The general liability rating engine implements the following basic strategy.

  1. The default rating engine determines which exposures are rate scalable. To do this, the default rating engine examines the class code on each exposure. If the class code of the exposure has a Basis whose Auditable property has the value false, then this exposure is rate scalable. Note that the Auditable flag is not directly on the class code or the exposure.
  2. The default rating engine rates each rate scalable exposure in sliced mode, just like a standard line of business rating engine iterates across slice dates.
  3. The default rating engine rates each basis scalable (non-rate-scalable) exposure in window mode, similar to the workers’ compensation rating engine.
  4. The default rating engine adds taxes and fee costs in window mode.

Another unusual aspect of general liability rating is that there could be 1, 2, or 4 costs for each coverage, depending on two different factors.

  • If the line has its SplitLimits property has the value true, then the rating engine tracks a bodily injury (PI) limit separate from property damage (PD) limit. Otherwise, the rating engine treats them as a single value called a combined single limit (CSL).
  • General liability coverages have cost sublines for both Premises and Products. The general liability rating lookup table for a limit returns two values: one for Premises and one for Products. If it returns a zero for either subline, the default rating engine throws away that cost information. the rating engine may return a non-zero cost for both sublines, or it may return a cost only for one of the two sublines.

Because of a combination of these two factors, each coverage on each exposure can generate 1 cost, 2 costs, or 4 costs.

Be very careful if you modify any code related to general liability or any new similar lines of business. You must ensure that your code tracks the proper number of costs (1, 2 or 4) as described in this topic.

For more details, refer to the implementation class gw.lob.gl.rating.GLRatingEngine.