1 / 25

Quantifying the Impact of Non-Modeled Catastrophes

Quantifying the Impact of Non-Modeled Catastrophes. Israel Krakowski Sara Drexler CAS Ratemaking Seminar March, 2003. Introduction. Methods to Estimate Catastrophe Provision from Historical Data Balance credibility and responsiveness issues

Télécharger la présentation

Quantifying the Impact of Non-Modeled Catastrophes

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. Quantifying the Impact of Non-Modeled Catastrophes Israel Krakowski Sara Drexler CAS Ratemaking Seminar March, 2003

  2. Introduction • Methods to Estimate Catastrophe Provision from Historical Data • Balance credibility and responsiveness issues • Appropriate provision for individual states as well as Countrywide • We are only going to discuss non-hurricane and non-earthquake catastrophe estimates • In this session we will cover • A brief overview of existing methodologies • Several new methodologies developed internally at Allstate by Israel Krakowski

  3. Agenda • Overview • Vocabulary • History of Methods • Description of New Methods • State Relativity Weighted with Countywide Complement • State Relativity Weighted with Regional Complement • Dual Capping Method • Summary/Discussion

  4. Overview • Increasing cat losses per AIY over time • “One in 100 year” events for a particular state occur relatively frequently on a countrywide basis • In response, we are researching new methods to estimate future expected catastrophe losses • This presentation shows methods to develop catastrophe provision, not how to incorporate provision into the indication

  5. Vocabulary • AIY – Amount of Insurance years 1AIY=$1,000 of dwelling coverage • Damage Ratios – Losses/AIY

  6. History of Methods

  7. Goals of New Methods • Develop an accurate, stable provision by state that results in an appropriate provision on a countrywide basis • Systematic approach to handle extreme events so a single outlying year does not drive the cat provision for a state • Appropriate application of credibility procedure • Provide result that is responsive to recent demographic and cat definition shifts • Method should be “sellable” to insurance departments and regional staff

  8. Method to be Presented • State relativity weighted with countrywide • State relativity weighted with regional • Refinement of Dean et. al. paper • Several other methods are included with Forum paper and will not be discussed here

  9. Data Considerations • Summary by state and calendar year of Allstate 1971-2000 catastrophe losses and AIY for HO, Renters, and Condo • Additional years have been added since the original research and are not included in this presentation • The nature of the analyses was constrained by the data available • From 1988 forward additional detail available: e.g. could have broken out catastrophes by peril

  10. State Relativity Weighted with Countrywide Complement – General Outline • Develop State Damage Ratios • Calculate Countywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision

  11. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Develop each state’s damage ratios for years 1981-2000 • State Damage Ratios – Losses/AIY • Only use years 1981 forward. Data for years 1971 through 1980 is sparse as evidenced by yearly variance

  12. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Each year’s Countrywide damage ratio is calculated as the weighted average of state damage ratios using latest year AIYs as weights • Countrywide catastrophe provision is the arithmetic average of recent years’ damage ratios • When breaking down the historical period into distinct segments the trend is flat for each period (Figure 1) • Using only most recent years allows us to be responsive to current level of catastrophe exposure • Can use a 10 year average or years since 1990 • One can tack on an additional load to be conservative

  13. Figure 1 YearsLinear trend 1971-1978 0.006 1979-1989 0.000 1990-1999 -0.019 1990-2000 -0.010

  14. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Calculate state relativities as the ratio of state damage ratios to countrywide damage ratios • Relativities should be more stable than damage ratios • Trend should not be a problem so we can use more years of data than the Countrywide Catastrophe Provision (Exhibit 1)

  15. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Any relativity greater than the mean plus three standard deviations is capped to the next lowest relativity (not the cap number) • Benefit of capping process • Represents a systematic approach to dealing with extreme events • Cap is dynamic and is allowed to shift if exposure in a state is changing over time • Censoring at the cap would not have much impact and therefore would not result in increased stability • Drawbacks of capping process • A year with a lower relativity could result in a higher catastrophe provision than a slightly higher one • A higher than average single year could lead to a less stable average relativity if it is also used as the “next lowest relativity” in the capping process

  16. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Calculate arithmetic average of 1981-2000 capped relativities • No benefit of weighting relativities has been shown since relationship of variability to exposure level is unclear • Arithmetic average relativity does not differ significantly from an AIY weighted average

  17. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Uses Buhlmann credibility factor: n/(n+k) • n = number of years of relativities in average • We use number of years rather than exposures because exposures not independent, especially past a certain threshold where exposure concentration increases • k = average process variance/variance of hypothetical means • The process variance and variance of hypothetical means are calculated using all available years of capped relativities across all states • Complement of credibility of 1.000 is not appropriate when there is a wide spread of average relativities (Exhibit 2)

  18. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • The unbalanced state relativities result in a countrywide relativity of less than 1.000. Relativities are adjusted: • Determined countrywide expected losses based on the countrywide selected catastrophe factor • Sum the pre-balanced expected losses across all states • Distribute the difference between 1 and 2 in proportion to each state’s standard deviation measured in latest year expected losses. • Using this approach has several benefits: • Results in an appropriate provision countrywide • It compensates for high (low) relativity states being underestimated (overestimated) by the use of a 1.000 complement of credibility • Resulting cat load is a function of each state’s size and variability • Exhibit 3 Shows numerical example

  19. Develop State Damage Ratios • Calculate Countrywide Damage Ratios • Calculate State Relativities • Cap State Relativities • Average Capped Relativities • Credibility Weight with CW Average of 1.000 • Balance Back to CW Average of 1.000 • Calculate Statewide Catastrophe Provision State Relativity Weighted with Countrywide Complement • Statewide catastrophe provision is calculated by multiplying capped, credibility weighted, balanced relativity by the countrywide catastrophe provision • Benefits of method: • Allows use of long term data to determine relativity while using more responsive data for countrywide provision • Adjustments to data are determined objectively with each state’s characteristics used to determine both capping and balancing • Drawbacks of method: • Complement of credibility is suspect • Despite the fact that any state has the potential to have capped losses, the capping/balancing process could be controversial

  20. State Relativity Weighted with Regional Complement • Regions are developed by combining several contiguous states with similar damage ratios and using correlations to determine close calls • Region-wide damage ratios are also calculated by weighting with latest year’s AIYs • Method credibility weights each state with its own region rather than countrywide

  21. State Relativity Weighted with Regional Complement • For regional methods credibility weight differs from countrywide in the expected process variance • Took weighting of state’s own sample process variance with average of all states in region’s process variance • Took weighting of maximum sample process variance for all states with average of all state’s sample process variance • Also adjusted VHP to eliminate negative values • As with previous method • State relativities are balanced to Regions which are themselves balanced to provide an adequate countrywide provision • We are using relativities which allows us to use all available valid years without concern for trend in damage ratios • Exhibit 4 shows example

  22. State Relativity Weighted with Regional Complement • Benefits • Complement of credibility is more appropriate than countrywide • We are able to consider the individual state variability (in our estimation of the process variance) in the credibility weight. • No need for capping since adjustments to credibility procedure neutralizes the impact of extreme events • Drawbacks • Regions are not optimal because they are developed around political boundaries when they theoretically should be developed around geographic boundaries • States may argue if extremely high cat state is included in their region

  23. Dual Capping Method • Method originally proposed by Dean et al • Rank catastrophe loss ratios over all available years (17 in Dean paper) • Censor losses above and below • Divide net excess losses by all years EP to get load • This is added to each individual year’s capped ratios • Problems • Historical premium inadequacies can distort • Historical catastrophe loss ratios are trended or not • Trending is problematic • Not trending is inconsistent with non catastrophe methodology and distorts which year will be capped • 17 years is not enough

  24. Dual Capping Method • Proposed modification • Rather than go back in time to get more points use a region for the latest 10 years. If region has, e.g., 7 states, there are 70 points • Rank all the damage ratios, rather than loss ratios, and double censor as before • Benefits • No premium adequacy problem • No trending problem • Contains more points • Drawbacks • Some states don’t fit in group well and will be overly capped • Looks too much like smoothing losses between states • Exhibit 5 shows example

  25. Summary • We’ve reviewed several methods to develop non-modeled catastrophe provisions • All of these methods attempt to balance the need for stable long term averages while being responsive to more recent catastrophe exposure distributions • What are your thoughts on the models presented or other alternatives? Use the following criteria to evaluate: • Accuracy • Stability • Sellability

More Related