List 5 estimates should be incorporated within methods when estimating unpaid claims, as suggested by B/S
1. Projections of reported claims2. Projections of paid claims3. Promections of ult rpt cc and severity4. Estimates of # and ave amt of o/s claims5. Claim ratio estimates
Whenever possible, what 3 types of analysis should be incorportated?
1. Credibility2. Regreassion analysis3. Data smoothing
It's the actu's responsibility to select the most appropriate estimate of unpaid claims. What should s/he select & what is gained from analysis?
Select highly stable method and highly responsive
Informative to see spread of estimates from different approaches * Helps with understanding the range and distn of possibilities* May show sensitivity of estimates to varying assumptions and various methodologies
List the 4 steps involved in the general selection process for an estimate of unpaid claims
1. Actu must select more appripriate estimate of unpaid claims* Select from highly stable methods and highly responsive
2. Retroactively test methods if sufficient claim history is available * Evaluate appropriateness of technique* Determine if accurate & unbiased
3. Actuary should explain significant differences btwn projections of various mtds* Changes in operations, procedures
4. Final check of selected ultimate claims* Include calculation of CR, Sev, Pure Prem, Claim Freq* Compared implied Case O/S & unreported claim on open and unrpted clms
When reviewing reasonableness of selected ult clms, what stats should be reviewed?
perspective of yr-to-yr changesknowledge gained from meetings w/ mgmtknowledge of industry in general
Name 3 acceptable ways to select ultimate claims
Select 1 method for all yearsSelect different methods for different AYsUse a weighted ave based on assigned weights to various mtds
Hugh White's question: You try to establish a reserve but find that the rpted % of expected losses is 8% higher than it should be. List 3 Options & their method names.
1. Reduce the Bulk Reserve (IBNR) a corresponding amt b/c you think it was an acceleration in the rate of reportEC Tech2. Leave the IBNR at the same % level of expected claims b/c you think there was a random fluctuation like a large claimBF Tech3. Increase the IBNR proportionatly to the increase in actual reported over expected rpted b/c not 100% sure of exp claim estimateDvmt Tech