Manipulating Correlations Through Latent Drivers

13 Pages Posted: 12 Nov 2010

See all articles by Jennifer Bender

Jennifer Bender

State Street Global Advisors

Jyh-Huei Lee

MSCI Inc.

Dan Stefek

MSCI Inc.

Date Written: May 21, 2010

Abstract

The ability to manipulate correlation matrices is useful for a number of applications in finance, including stress testing and portfolio construction. We outline a flexible framework that enables managers to manipulate correlations based on their beliefs and preserve the positive semi-definiteness of the matrix. The framework links the returns for which correlations are captured in the matrix through unobservable factors or drivers, called latent drivers. Changes in correlations can then be introduced through these drivers. Users can specify levels of exposure to a driver based on their prior beliefs. The framework can be applied to any correlation matrix - asset, factor, or asset class.

Keywords: Manipulate Correlation Matrices, Stress Testing Portfolio, Construction Latent Drivers, Flexible Framework, Positive Semi-Definiteness Matrix, Unobservable Factors Drivers

Suggested Citation

Bender, Jennifer and Lee, Jyh-Huei and Stefek, Dan, Manipulating Correlations Through Latent Drivers (May 21, 2010). MSCI Barra Research Paper No. 2010-20, Available at SSRN: https://ssrn.com/abstract=1707506 or http://dx.doi.org/10.2139/ssrn.1707506

Jennifer Bender (Contact Author)

State Street Global Advisors ( email )

1 Lincoln Street
28th Floor
Boston, MA 02111
United States

Jyh-Huei Lee

MSCI Inc. ( email )

88 Pine Street
2nd Floor
New York, NY 10005
United States

Dan Stefek

MSCI Inc. ( email )

88 Pine Street
2nd Floor
New York, NY 10005
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
328
Abstract Views
1,553
Rank
180,633
PlumX Metrics