Tree on paths method and system for modeling securities

• US 7,818,244 B1
• Filed: 12/20/2002
• Issued: 10/19/2010
• Est. Priority Date: 12/20/2002
• Status: Active Grant

• Current Assignee: Fannie Mae
• Original Assignee: Fannie Mae
• Inventors: Shulman, Alexander L., Dharan, Venkat G., Kramin, Marat V.
• Primary Examiner(s): Kazimi; Hani M.
• Assistant Examiner(s): GRAHAM, CLEMENT B
• Application Number:
  US10/324,100
• Time in Patent Office: 2,860 Days
• Field of Search: 705/50, 705 35- 38, 705/47, 705/8
• US Class Current: 705/37
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

A system for implementing a tree on paths method for generating a recombining interest rate tree from previously generated paths. The inventive method is a tree on paths method whereby a tree is created from previously generated paths. The tree on paths method uses Monte Carlo simulations to generate paths and uses a recombining algorithm to obtain a computationally efficient tree from the generated paths.

Claims

1. A electronic programmed computer implemented method for approximating distribution of a Stochastic process for modeling derivative securities, the method comprising the steps of:

specifying a stochastic process for an underlying state variable;

using Monte Carlo simulations, by a computer, to generate a plurality of paths corresponding to the underlying state variable, each of the plurality of paths comprising states that represent future evolutions of the underlying state variable for a given time period through a given maturity date;

generating a tree using a computer, based upon the plurality of paths, the tree comprising a recombining of states of the plurality of paths at each of a plurality of time layers, wherein the recombining for a given time layer determines a set of closest generated states for the given time layer as compared to a previous time layer, and applies a transition probability to link the set of closest generated states for the given time layer to a previous state; and

wherein the underlying state variable is a primary state variable having a corresponding secondary state variable, and wherein specifying the stochastic process, using the Monte Carlo simulations, and generating the tree implement the primary state variable and the secondary state variable and using the generated tree to evaluate derivative securities.

2. The method of claim 1 wherein the Monte Carlo simulations use a Euler difference equation.

3. The method of claim 1, further comprising the step of structuring the states in the plurality of paths prior to generating the tree.

4. The method of claim 3, wherein structuring the states comprises sorting the states in increasing order to assist the recombining.

5. The method of claim 3, wherein structuring the states comprises sorting the states in decreasing order to assist the recombining.

6. The method of claim 1, wherein the tree is used to value different types of derivative securities.

7. The method of claim 1, wherein the underlying state variable is an interest rate.

8. The method of claim 1, wherein the underlying state variable is used to determine an interest rate.

9. A non transitory computer-readable medium having program code stored therein for approximating distribution of a stochastic process in modeling derivative securities, the program code including instructions which, when executed by a computer, perform operations comprising:

specifying a stochastic process for an underlying state variable;

using Monte Carlo simulations, by a computer, to generate a plurality of paths corresponding to the underlying state variable, each of the plurality of paths comprising states that represent future evolutions of the underlying state variable for a given time period;

generating a tree, using a computer, based upon the plurality of paths, the tree comprising a recombining of states of the plurality of paths at each of a plurality of time layers, wherein the recombining for a given time layer determines a set of closest generated states for the given time layer as compared to a previous time layer, and applies a transition probability to link the set of closest generated states for the given time layer to a previous state; and

wherein the underlying state variable is a primary state variable having a corresponding secondary state variable, and wherein specifying the stochastic process, generating the Monte Carlo simulations, and generating the tree implement the primary state variable and the secondary state variable and using the tree to evaluate the derivative security.

10. The non transitory computer readable medium of claim 9, the underlying state variable is used to determine an interest rate.

11. The non transitory computer readable medium of claim 9, the underlying state variable is used to determine an interest rate.