1 / 18

Estimating the Parameter Risk of a Loss Ratio Distribution

Estimating the Parameter Risk of a Loss Ratio Distribution. Chuck Van Kampen, FCAS American Agricultural Ins. Co. CaRe Seminar Philadelphia, Pa June 2, 2003. Motivation For Study . Price Stop Loss Reinsurance Contracts DFA Analysis. Base Case Data. Pricing Question. Assume that:

long
Télécharger la présentation

Estimating the Parameter Risk of a Loss Ratio Distribution

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Estimating the Parameter Risk of a Loss Ratio Distribution Chuck Van Kampen, FCAS American Agricultural Ins. Co. CaRe Seminar Philadelphia, Pa June 2, 2003

  2. Motivation For Study • Price Stop Loss Reinsurance Contracts • DFA Analysis

  3. Base Case Data

  4. Pricing Question • Assume that: • The LogNormal Distribution is the correct model distribution • That the parameters are known with absolute certainty • How well does ten years of data predict future experience?

  5. Bootstrap with Simulations • Use Parameters to simulate 10,000 ten-year blocks of loss ratios • How many simulated ten-year blocks have a mean, standard deviation and skewness close to the actual data?

  6. What is Close? • Ranges around actual data: • Mean 63.52% to 64.77% • Std Dev .0662 to .0762 • Skewness .29 to .71

  7. How Well Does Ten Years Predict Future Experience? • 117 out of 10,000 simulations (1.17%) have the mean, standard deviation and skewness close to the actual data • Conclude it is unlikely that the actual ten-year block of data will provide parameters that are close to the true underlying distribution

  8. Alternative Pricing • Don’t Use Best Fit parameters • Instead determine: • What sets of parameters could have produced the actual data • And the relative probability or each of these parameter sets

  9. Determining Viable Parameter Sets • Use a macro to step through parameter ranges • Create 10,000 ten-year blocks for each parameter set • Count ten-year blocks that have mean, std dev and skew close to actual data for each parameter set

  10. Sample of Parameter Sets

  11. Parameter Set Relative Probabilities – Side View

  12. Parameter Set Relative Probabilities – Top View

  13. Comparison of Parameter Set and Fitted Expected Loss

  14. Sensitivity Testing • Increase the mean • Increase the standard deviation • Increase the skew • Fewer years of data

  15. Load by ELOL for Excess Layers

  16. Loads For Primary Loss Ratios

  17. Comments on Determining Viable Parameter Sets • Step size through the parameter set ranges • Size of parameter set ranges • What is close to the actual data • Criteria used

  18. Considerations • Simulation is used, not exact • Cat exposures should be removed • Requires a fair amount of judgment • Exposures not present in data are not taken into account • Process Risk still present and not accounted for in this methodology

More Related