Method and system for evaluating insurance liabilities using stochastic modeling and sampling techniques

• US 8,126,747 B2
• Filed: 01/20/2011
• Issued: 02/28/2012
• Est. Priority Date: 09/23/2009
• Status: Active Grant

• Current Assignee: Towers Watson Delaware Incorporated (Willis Towers Watson plc)
• Original Assignee: Watson Wyatt & Company
• Inventors: Vadiveloo, Jeyaraj
• Primary Examiner(s): Akintola, Olabode
• Application Number:
  US13/009,914
• Publication Number:
  US 20110106572A1
• Time in Patent Office: 404 Days
• Field of Search: 705/4, 705 35- 40
• US Class Current: 705/4
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

In computer-implemented methods and systems for estimating financial modeling outcomes, financial data segmented into a number (x) of classes and scenario data for a set of model scenarios are processed to obtain an estimated model outcome distribution. The class segments are mutually exclusive and collectively exhaustive of the financial data. Multiple model tests are performed with samples of the financial data until a cumulative model outcome distribution is within a pre-determined acceptable tolerance limit from a distribution of fully assessed model outcomes obtainable by performing a single test of the scenarios using all of the financial data. The number (x) of classes, the sample size (z), and a number (y) of times that the tests are performed ensure that the cumulative model outcome distribution is within the pre-determined acceptable tolerance limit from the distribution of fully assessed model outcomes.

Claims

1. A computer-implemented method for performing analysis of financial data, comprising the steps of:

storing, in a computer readable memory,

financial data related to a population of financial data records and segmented into a number (x) of classes, wherein the class segments are mutually exclusive and collectively exhaustive of the financial data, and

scenario data for a set of scenarios, wherein each of the scenarios is defined at least in part by a set of variables, and wherein the scenario data for each of the scenarios comprise at least some parameter values for the set of variables;

providing a computer processor associated with the computer readable memory with a model of a system defined at least in part by the set of variables; and

processing, with the computer processor, the financial data and the scenario data using the model to obtain an estimated model outcome distribution comprising a distribution of estimated model outcomes relating to the system and based on the set of scenarios, the processing further comprising;

selecting a first subset of the financial data as a first sample, the first subset drawn without replacement from each of the class segments of the population and having a sample size (z);

performing, with the model and the first subset, a first test of the set of scenarios to obtain a first set of sample outcomes for each of the scenarios; and

repeating the selecting and performing steps using additional subsets of the financial data to perform additional tests of the set of scenarios and to obtain additional sets of sample outcomes for each of the scenarios;

wherein the additional subsets are drawn without replacement from each of the class segments;

wherein the first sample outcomes and the additional sample outcomes are combined to create a cumulative estimated model outcome distribution;

wherein the selecting and performing steps are repeated until the cumulative model outcome distribution is within a pre-determined acceptable tolerance limit from a distribution of fully assessed model outcomes for the set of scenarios obtainable by performing, with the model, a single test of the set of scenarios using all of the data; and

wherein the cumulative model outcome distribution is identified as the estimated model outcome distribution;

wherein the number (x) of classes, the sample size (z), and a number (y) of tests comprising a count of the number of times that the performing step is conducted ensure that the cumulative model outcome distribution is within the pre-determined acceptable tolerance limit from the distribution of fully assessed model outcomes.

2. The computer-implemented method of claim 1, further comprising using stratified sampling techniques to generate the samples.

3. The computer-implemented method of claim 1, wherein selecting the first sample and the additional samples comprises selecting, from a selected class segment, a number of records that is proportional to a count of records in the selected class segment relative to a total count of records in the financial data.

4. The computer-implemented method of claim 1, wherein the additional subsets are drawn with replacement from the population and without replacement from each of the class segments after prior samples are returned to the population.

5. The computer-implemented method of claim 1, wherein the additional subsets have the sample size (z).

6. The computer-implemented method of claim 1, further comprising segmenting the financial data into the selected number (x) of classes before storing the financial data in the computer readable memory.

7. The computer-implemented method of claim 1,

wherein the financial data comprises insurance policyholder data;

wherein the model comprises a model of insurance liability; and

wherein the model outcomes comprise measures of a risk comprising a likelihood of an occurrence of the insurance liability for a scenario in the set of scenarios.

8. The computer-implemented method of claim 1, further comprising

randomly sampling the financial data from each of the class segments to create a pseudo-population of records;

wherein processing the financial data and the scenario data further comprises treating the pseudo-population as the population for the selecting step and the performing step to obtain pseudo-population outcomes for each of the scenarios; and

wherein the combining step further comprises combining the pseudo-population outcomes to obtain a distribution of cumulative model outcomes for the pseudo-population.

9. The computer-implemented method of claim 8, further comprising repeating the selecting and performing steps using the number (x) of classes, the sample size (z), and the number (y) of tests to estimate the fully assessed model outcomes for the entire population.

10. The computer-implemented process of claim 8, wherein each of the pseudo-population outcomes comprises a value;

further comprising

adjusting each of the pseudo-population outcomes relative to the population so that a value of a selected pseudo-population outcome associated with a selected scenario comprises a preliminary estimate of a value of a fully assessed model outcome associated with the selected scenario and obtainable by performing, with the model, the single simulation of the scenario using the population of records.

11. The computer-implemented method of claim 1, wherein combining the sample outcome distributions further comprises averaging sample outcomes associated with each of the scenarios for each of the tests to obtain an average estimated model outcome for each of the scenarios.

12. The computer-implemented method of claim 1, further comprising ordering the estimated model outcomes to obtain percentiles in the estimated model outcomes distribution.

13. The computer-implemented method of claim 1,

wherein each of the sample outcomes comprises a value;

and

wherein combining the sample outcomes further comprises adjusting each of the sample outcomes relative to an entire dataset of the financial data so that the value of a selected sample outcome associated with a selected scenario comprises a preliminary estimate of the value of a fully assessed model outcome associated with the selected scenario and obtainable by performing, with the model, a single simulation of the scenario using the entire dataset.

14. The computer-implemented method of claim 1,

wherein the set of scenarios comprises a baseline set of scenarios, the estimated model outcomes comprise a baseline estimated model outcomes, and the estimated model outcome distribution comprises a baseline estimated model outcome distribution;

and

wherein the method further comprises;

storing, in the computer readable memory, second scenario data for a second set of scenarios;

performing the processing step with the financial data and the second scenario data using the model to obtain a second estimated model outcome distribution associated with the second set of scenarios; and

comparing the baseline estimated model outcomes with the second estimated model outcomes in order to identify differences therebetween.

15. The computer-implemented method of claim 14, wherein the comparing step comprises determining a ratio of percentile values between the second estimated model outcome distribution and the baseline estimated model outcome distribution.

16. The computer-implemented method of claim 14, wherein the second estimated model outcome distribution is developed using the same set of samples used to develop the baseline estimated model outcome distribution.

17. The computer-implemented method of claim 1, further comprising selecting the number (y) of tests and the sample size (z), by:

performing the processing step with a plurality of combinations of the sample size and the number of tests to identify an optimal combination of the sample size and the number of tests at which is obtained an optimal speed of convergence of the cumulative model outcome distribution over repetitions of the selecting and performing steps; and

selecting the optimal combination as the number (y) of tests and the sample size (z).

18. The computer-implemented method of claim 17, further comprising:

selecting the sample size (z) and a plurality of number of tests;

performing the processing step with the plurality of number of tests to identify an optimal number of tests required to obtain the optimal speed of convergence; and

selecting the optimal number of tests and the sample size (z) as the optimal combination.

19. The computer-implemented method of claim 17, further comprising selecting the number (y) of tests and a plurality of sample sizes;

performing the processing step with the plurality of sample sizes to identify an optimal sample size required to obtain the optimal speed of convergence; and

selecting the optimal sample size and the number of tests (y) as the optimal combination.

20. The computer-implemented method of claim 1, wherein the samples are representative of the classes in the financial data.