Method and system for managing collateralized obligations to satisfy predetermined investment ratings requirements
A method and system are provided which generate a collateral quality formula, based on a data set of collateral quality values, used to guide purchases and sales of collateral by CDO managers to satisfy investment ratings requirements, such as those established by independent ratings services, such as, for example, Moody's Investors Service, without having to adhere to the more limited defined parameters of a collateral quality value matrix.
1. A computer implemented method for managing a collateralized obligation by the computer to satisfy investment ratings requirements, comprising the steps of:
determining, by the computer, at least one combination of values of quality data factors associated with collateral to satisfy requirements for a preselected investment ratings category for a collateralized obligation including said collateral, said collateral quality data factors comprising a diversity score, a minimum average recovery rate, a weighted average rating factor, and a weighted average margin, said at least one combination of values of collateral quality data factors forming a basis for a matrix of values of collateral quality data factors;
expanding, by the computer, said matrix to a size mathematically suitable for performing a correlation/regression analysis on said values of collateral quality data factors;
performing, by the computer, a correlation/regression analysis on said values of collateral quality data factors to generate a mathematical relationship that best fits said values of collateral quality data factors; wherein said mathematical relationship comprises a plurality of tests for determining a necessary range of values for each of said collateral quality data factors to satisfy said requirements for said preselected investment ratings category; and
using, by the computer, said mathematical relationship in lieu of said matrix to guide at least one of purchases and sales of said collateral such that said collateralized obligation satisfies said requirements for said preselected investment ratings category.
2. The method according to claim 1 , further comprising the steps of testing said mathematical relationship for error, and when an error is present optimizing results of said correlation/regression analysis by statistical fitting.
3. The method according to claim 2 , wherein said statistical fitting is effected using at least one of a gradient descent algorithm, the Gauss-Newton algorithm, the Levenberg-Marquardt algorithm, and a probabilistic algorithm.
4. A system for managing a collateralized obligation to satisfy investment ratings requirements comprising:
a computer configured to determine at least one combination of values of quality data factors associated with collateral to satisfy requirements for a preselected investment ratings category for a collateralized obligation including said collateral, said collateral quality data factors comprising a diversity score, a minimum average recovery rate, a weighted average rating factor, and a weighted average margin, said at least one combination of values of collateral quality data factors forming a basis for a matrix of values of collateral quality data factors;
the computer configured to expand said matrix to a size mathematically suitable for performing a correlation/regression analysis on said values of collateral quality data factors;
the computer configured to perform a correlation/regression analysis on said values of collateral quality data factors to generate a mathematical relationship that best fits said values of collateral quality data factors, wherein said mathematical relationship comprises a plurality of tests for determining a necessary range of values for each of said collateral quality data factors to satisfy said requirements for said preselected investment ratings category; and
the computer configured to guide at least one of purchases and sales of said collateral using said mathematical relationship in lieu of said matrix such that said collateralized obligation satisfies said requirements for said preselected investment ratings category.
5. The system according to claim 4 , wherein the computer is further configured to test said mathematical relationship for errors and optimize results of said correlation/regression analysis by statistical fitting when an error is present.
6. The system according to claim 5 , wherein said statistical fitting is effected using at least one of a gradient descent algorithm, the Gauss-Newton algorithm, the Levenberg-Marquardt algorithm, and a probabilistic algorithm.
7. A computer usable medium having a plurality of instructions, the plurality of instructions including instructions which, when executed by a processor, cause the processor to perform the steps of a method for managing a collateralized obligation to satisfy investment ratings requirements, said method comprising the steps of:
determining at least one combination of values of quality data factors associated with collateral to satisfy requirements for a preselected investment ratings category for a collateralized obligation including said collateral, said collateral quality data factors comprising a diversity score, a minimum average recovery rate, a weighted average rating factor, and a weighted average margin, said at least one combination of values of collateral quality data factors forming a basis for a matrix of values of collateral quality data factors;
expanding said matrix to a size mathematically suitable for performing a correlation/regression analysis on said values of collateral quality data factors;
performing a correlation/regression analysis on said values of collateral quality data factors to generate a mathematical relationship that best fits said values of collateral quality data factors; wherein said mathematical relationship comprises a plurality of tests for determining a necessary range of values for each of said collateral quality data factors to satisfy said requirements for said preselected investment ratings category; and
guiding at least one of purchases and sales of said collateral using said mathematical relationship in lieu of said matrix such that said collateralized obligation satisfies said requirements for said preselected investment ratings category.
8. The computer usable medium according to claim 7 , further comprising means for testing said mathematical relationship for error, and means for optimizing results of said correlation/regression analysis by statistical fitting when an error is present.
9. The computer usable medium according to claim 8 , wherein said statistical fitting is effected using at least one of a gradient descent algorithm, the Gauss-Newton algorithm, the Levenberg-Marquardt algorithm, and a probabilistic algorithm.