Dimensionality, performance, and stress testing of multifactor equity models

A global multi-factor equity model had been previously developed. It employs a group of sector indices for different regions using Dow Jones, MSCI, and indices of local stock exchanges. PCA is used to produce independent factor variables, and then specific equities are regressed against the PCA factors. By transforming factors back into real world variables (indices) stress tests against movements in the original indices show their impact in a global equity portfolio. One question to be addressed is how many PCA dimensions should be selected for each equity. A systematic study of the performance of the methodology is needed, and potential changes to the methodology explored. The methodology will be extended to ETFs and mutual funds to better measure and manage investors? risks. A systematic study will be undertaken to analyze and measure the performance, stability, and risk management utility of the approach.

Faculty Supervisor:

David Lozinski

Student:

Lingting Meng

Partner:

RiskGrid Technologies Inc.

Discipline:

Mathematics

Sector:

University:

McMaster University

Program:

Accelerate

Current openings

Find the perfect opportunity to put your academic skills and knowledge into practice!

Find Projects