RBC R4 "C" Question
So I know for r5 to calculate company LR we'd average the last 10 years if we're given a table of historical loss ratios, capped at 300% for any given year, when calculating C in the risk charge factor.
Is there an equivalent "capping" rule for LDFs C for r4?
Comments
Odomirok page 274 suggests 400%:
Company “development factor”
The reporting entity’s own loss experience is considered by adjusting the industry reserve RBC percent by the company “development factor” by line of business. This development factor is calculated as the ratio of the sum of incurred loss and DCC from nine prior accident years evaluated as of the current year to the sum of the initial evaluations of those incurred amounts. The current incurred loss and DCC values come from Schedule P, Part 2, column 10, with the initial values coming from the first incurred value shown for each accident year. The initial values lie along the diagonal. This development factor measures how the initial estimates of ultimate loss and DCC have developed based on what the company currently knows. The factor is capped at 400% to limit the impact of anomalous, one-time results.
That's correct, and just to emphasize, the 300% LR cap for R5 is applied to individual years, whereas the 400% LDF cap is applied to the aggregate LDF.
(You can think of the company LR as being a straight average of 10 individual years, whereas the company LDF is a weighted average for the last 9 years. Note that you can't use 10 years because the most recent year hasn't had a chance to develop.)
Thanks Paul and Graham.
So to use the Schedule P Part 2 from the Liberty Mutual's 2016 annual statement in the Schedule P section, would the company LDF be calculated as the sum of column 10 (ignoring row 11) divided by the sum of the diagonal (again, ignoring row 11)? And the maximum this value can be is 1.4?
Edit: Also excluding prior row in addition to excluding row 11
Right, use the 2007 row to the 2015 row. So the numerator would be the sum of these values in Column (10):
and the denominator would be the sum of this diagonal:
But the result would be capped at 4.0 (400%) rather than 1.4. The maximum of 400% does seem high, but 1.4 seems too low, even using a 9-year average, if it's a long-tailed line.
And since we're on the topic, there are situations where the company may not use its own data. Even though the CAS likes to pick questions from bullet point lists, my feeling is that these exceptions are too detailed.
thanks for clarifying!