System, method and framework for generating scenarios

• US 7,599,870 B2
• Filed: 04/12/2002
• Issued: 10/06/2009
• Est. Priority Date: 04/12/2002
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: Glo Software, LLC
• Inventors: Merkoulovitch, Leonid, Mutnikas, Yaacov, Rosen, Dan, Reynolds, Diane
• Primary Examiner(s): Shah; Kamini S
• Assistant Examiner(s): Gebresilassie; Kibrom K
• Application Number:
  US10/120,795
• Publication Number:
  US 20030195830A1
• Time in Patent Office: 2,734 Days
• Field of Search: 703/2, 705 35- 38, 705/10, 717/101
• US Class Current: 705/36.00R
• CPC Class Codes: G06Q 10/0635 Risk analysis of enterprise or organisation activities
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

System, method and framework for generating scenarios used in risk management applications. The present invention is based on a generic framework that provides levels of abstraction, segregates risk factors and models, and structures a scenario generation process. In one aspect of the present invention, there is provided a framework for scenario generation for use in a risk management application, where the framework defines a plurality of components associated with a scenario set, where each component is represented by at least one of a set of data structures, and where the set of data structures comprises: at least one first data structure defining a group of risk factors with similar statistical properties; at least one second data structure defining the future distribution or evolutionary process of a risk factor in the group of risk factors; a third data structure defining a calibrated model for generating scenarios, where relationships between risk factors of the group of risk factors are defined therein, and where the calibrated model associates each second data structure with a first data structure; and a fourth data structure specifying how the first, second, and third data structures are to apply to a user-specified risk management problem.

Claims

1. A computer-implemented method of risk factor scenario generation for use in a risk management application comprising:

receiving user input at a scenario builder graphical user interface, wherein the user input identifies a plurality of risk factors affecting a portfolio under consideration;

wherein each of said plurality of risk factors is an observable economic variable whose value or change in value translates into a change in the value of a portfolio under consideration, and wherein each of said plurality of risk factors is assigned to a risk factor class;

creating a plurality of Block data structures;

storing on a database said plurality of risk factors in said plurality of Block data structures such that each risk factor is stored in exactly one of the plurality of Block data structures, and such that risk factors stored in each Block data structure have similar statistical properties wherein risk factors with similar statistical properties are assigned to the same risk factor class;

creating at least one Model data structure, each Model data structure comprising data defining how values of the risk factors with similar statistical properties change;

building a Scenario Generator data structure, said Scenario Generator data structure comprising

a calibrated model for generating risk factor scenarios;

the calibrated model comprising

a list of blocks and models associating a Model data structure with each of said plurality of Block data structures; and

a codependent structure defining relationships between risk factors of said plurality of risk factors;

creating a Scenario Set Definition data structure specifying how said Scenario Generator data structure is to apply to a user-specified risk management problem;

generating a scenario set for said plurality of risk factors at a scenario engine server running on a machine to generate the scenario set by applying said Scenario Generator data structure, as specified by said Scenario Set Definition data structure, to a sampling of random numbers, wherein said generated scenario set comprises values of said plurality of risk factors at one or more future points in time; and

producing one or more data files comprising data associated with said generated scenario set, for computing a monetary value associated with the portfolio under consideration.

2. The method as claimed in claim 1, wherein said creating the plurality of Block data structures comprises defining a list of risk factors.

3. The method as claimed in claim 2, wherein said creating the plurality of Block data structures further comprises defining at least one of a preprocessing routine or a transformation technique.

4. The method as claimed in claim 1, wherein said creating the at least one Model data structure comprises defining a simulation equation and at least one calibration method.

5. The method as claimed in claim 1, wherein said building the Scenario Generator data structure comprises defining a list of associated groups of risk factors and future distributions or evolution processes, and defining a codependent structure in which relationships between said associated groups of risk factors and future distributions or evolution processes are defined.

6. The method as claimed in claim 5, wherein said building the Scenario Generator data structure further comprises defining a sampling method.

7. The method as claimed in claim 1, wherein said creating the Scenario Set Definition data structure comprises defining an identifier of said calibrated model and a plurality of structural parameters.

8. The method as claimed in claim 7, wherein said creating the Scenario Set Definition data structure further comprises defining a written description of said generated scenario set.

9. The method as claimed in claim 1, wherein each Model data structure defines, for each risk factor defined in the Block data structure with which the Model data structure is associated, how the value of said each risk factor changes with time.

10. The method as claimed in claim 1, wherein each Model data structure defines, for each risk factor defined in the Block data structure with which the Model data structure is associated, how the value of said each risk factor changes at a step.