Computer-implemented risk evaluation systems and methods

• US 7,610,257 B1
• Filed: 01/10/2007
• Issued: 10/27/2009
• Est. Priority Date: 01/10/2006
• Status: Active Grant

• Current Assignee: SAS Institute Incorporated
• Original Assignee: SAS Institute Incorporated
• Inventors: Abrahams, Clark Richard
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US11/651,914
• Time in Patent Office: 1,021 Days
• Field of Search: 706/62, 705/4, 705/38, 705/39
• US Class Current: 706/62
• CPC Class Codes: G06Q 20/10 specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

Computer-implemented systems and methods are disclosed herein that relate to processes, which construct an empirically derived and statistically based risk evaluation and policy formulation system. For example, a process can be configured so as to accept as input an information base in computer readable form and produce either a single or multistage system composed of alternative decision making strategies.

Claims

1. A computer-implemented system for evaluating risks associated with alternative strategies for assessing an entity with respect to a predetermined objective, comprising:

an action table;

wherein the action table contains a plurality of possible actions that can be taken with respect to the predetermined objective for various entity profiles;

wherein the action table contains an action for each entity profile;

one or more statistical data stores that are configured to contain risk amounts associated with each entity profile;

wherein the one or more statistical data stores are configured to contain inferred percentage distribution of applicants associated with each entity profile;

one or more profile identification data stores;

wherein a profile identification data store contains entity identification information for use in determining an action for the entity;

wherein an action is determined for the entity by comparing characteristics associated with the entity with characteristics associated with the entity profiles contained in the action table;

one or more improvement data stores;

wherein an improvement data store provides an indication in the improvement in risk based upon a change in one or more characteristics for an entity profile;

one or more protected class data stores;

wherein a protected class data store provides an indication of the impact of a policy change on a particular protected class, both with respect to the approval, collection, or promotion decision and relative to one or more characteristics, or combinations of characteristics, for an entity profile.

2. The system of claim 1, wherein a plurality of action tables is used to analyze a plurality of alternative policies;

wherein an action table is associated with a policy.

3. The system of claim 2, wherein historical data is used to generate an initial set of alternative policies.

4. The system of claim 1, wherein the statistical data stores include one or more data stores that contain percentage distribution forecasts for the entity profiles;

wherein the statistical data stores are used in generating a score value that indicates a level of risk associated with an action for the entity under evaluation.

5. The system of claim 1, wherein the possible actions relate to a credit approval process.

6. The system of claim 5, wherein the possible actions indicate accept or decline decisions with respect to the credit approval process.

7. The system of claim 6, wherein the action table is used within a multi-stage system to determine whether an entity is to be approved in the credit approval process.

8. The system of claim 7, wherein the first stage contains different actions than the actions in the second stage.

9. The system of claim 8, wherein the first stage includes accept or decline or investigate actions;

wherein the second stage is used if the decision from the first stage is an investigate action.

10. The system of claim 5, wherein the action table is used within a single-stage system to determine whether an entity is to be approved in the credit approval process.

11. The system of claim 1, wherein the statistical data stores include a forecast of how the total population of entities is distributed with respect to the characteristics, individually and in every possible combination of each with the others that are used to define an entity profile.

12. The system of claim 1, wherein significant shifts in the applicant population percentage distribution with respect to any profile, or group of profiles, can be identified and incorporated to maintain the system'"'"'s approval rate accuracy by a simple tally of the number of applicants corresponding to each cell in the action table and by adding the result to the pre-existing action table applicant cell count.

13. The system of claim 1, wherein significant shifts in the degree of risk that the system associates with each applicant profile specified in the action table can be identified and incorporated to maintain the system'"'"'s risk evaluation accuracy by a simple tally of the number of good and bad loans associated with each action table reference number that were booked over a specified time period in question and by adding the resulting tallies, respectively, to the pre-existing action table good and bad counts associated with each action table cell.

14. The system of claim 1, wherein significant shifts in the account population percentage distribution with respect to any behavioral profile, or group of behavioral profiles, can be identified and incorporated to maintain the system'"'"'s account mix by a simple tally of the number of accounts corresponding to each cell in the action table and by adding the results to the pre-existing action table respective cell count.

15. The system of claim 1, wherein significant shifts in the degree of risk that the system associates with each account profile specified in the action table can be identified and incorporated to maintain the system'"'"'s risk evaluation accuracy by a simple tally of the number of redeemed, and non-redeemed, accounts associated with each action table reference number over a specified time period and by adding the resulting tallies, respectively, to the pre-existing action table redeemed and non-redeemed counts associated with each action table cell.

16. The system of claim 1, wherein significant differences between actual versus estimated rates of account redemption relative to every account profile specified in the action table can be identified and incorporated to ensure that the system'"'"'s predicted account performance is sufficiently accurate for a collector planning a collection approach to an account, or accounts having identical behavior profiles, or credit personnel in evaluating a credit transaction authorization request or setting credit limits, or account evaluation programs at renewal time.

17. The system of claim 2, wherein two main account promotion objectives are specified,

wherein the objectives include response to an offer and subsequent high purchase activity.

18. The system of claim 2, wherein an account activation objective is specified.

19. The system of claim 1, wherein individual components of an underwriting or pricing policy, and combinations thereof, can be tested for statistical significance relative to a plurality of protected class definitions and other givens;

wherein the class definitions are selected from the group including race, ethnicity, gender, and age;

wherein the givens are selected from the group including loan amount, loan type, dwelling type, occupancy status, lien status, loan amount, borrower income, geographic market, channel, and terms and conditions of the loan agreement.

20. The system of claim 2, wherein expert or policy-based business rules are used to specify an initial set of mortgage company retail-telemortgage channel underwriting policies for conventional, owner-occupied, single-family dwelling, purchase mortgages, wherein the first action table is comprised of primary factors and the second stage action table is a collection of tables comprised of secondary factors conditional on combinations of the primary factors and the possible outcomes include specific counter-offers or alternative product or program offerings, and a specific pricing sheet with up-down adjustments based upon a set of factors that may contain additional variables not already considered in the decision process in addition to approve or decline.

21. The system of claim 20, wherein expert or policy-based business rules are used to specify an initial set of mortgage company retail-telemortgage channel underwriting policies for conventional, owner-occupied, single-family dwelling, re-finance mortgages where the current mortgage is with the same institution, or as a separate case where the current mortgage is with a different institution.

22. The system of claim 20, wherein expert and policy-based business rules are used to specify an initial set of mortgage company retail-telemortgage channel underwriting policies for conventional, owner-occupied, single-family dwelling, home improvement second lien mortgages where the value of the property financed is “

as-is”

, or as a separate case where the value of the property financed is “

after improvements”

.

23. The system of claim 1, wherein the possible actions relate to a credit promotion process.

24. The system of claim 23, wherein the possible actions indicate accept or decline decisions with respect to the credit promotion process.

25. The system of claim 24, wherein the action table is used within a multi-stage system to determine whether an entity is to be approved in the credit promotion process.

26. The system of claim 23, wherein the action table is used within a single-stage system to determine whether an entity is to be approved in the credit promotion process.

27. The system of claim 1, wherein the possible actions relate to an account management process.

28. The system of claim 27, wherein the possible actions indicate accept or decline decisions with respect to the account management process.

29. The system of claim 28, wherein the action table is used within a multi-stage system to determine whether an entity is to be approved in the account management process.

30. The system of claim 27, wherein the action table is used within a single-stage system to determine whether an entity is to be approved in the account management process.

31. The system of claim 1, wherein the system updates for shifts in applicant mix across action table cells and for shifts in the risk associated with those cells.

32. The system of claim 1, wherein the possible actions relate to a credit pricing process.

33. The system of claim 1, wherein the possible actions relate to an insurance policy underwriting process;

wherein types of insurance policies include automobile, renters hazard and theft, homeowners, income continuation, life, accident, and umbrella coverage.

34. The system of claim 1, wherein the possible actions relate to a pricing process for an insurance policy;

wherein types of insurance policies include automobile, renters hazard and theft, homeowners, income continuation, life, accident, and umbrella coverage.

35. A computer-implemented method for evaluating risks associated with alternative strategies for assessing an entity with respect to a predetermined objective, comprising:

determining an action for the entity by comparing characteristics associated with the entity with characteristics associated with entity profiles contained in an action table;

wherein the action table contains a plurality of possible actions that can be taken with respect to the predetermined objective for various entity profiles;

wherein the action table contains an action for each entity profile;

wherein one or more statistical data stores are configured to contain risk amounts associated with each entity profile;

wherein the one or more statistical data stores are configured to contain inferred percentage distribution of applicants associated with each entity profile;

wherein a profile identification data store contains entity identification information for use in determining an action for the entity;

using an improvement data store to provide an indication in the improvement in risk based upon a change in one or more characteristics for an entity profile; and

using a protected class data store to provide an indication of the impact of a policy change on a particular protected class, both with respect to the approval, collection, or promotion decision and relative to one or more characteristics, or combinations of characteristics, for an entity profile.