Rufus Rankin is the director of research and a portfolio manager at Equinox Institutional Asset Management. Equinox is a comprehensive alternative investment provider with multiple offerings across global managed futures, commodity, and hedge equity asset classes. He is frequently featured in leading industry publications and is the author of Multi-Dimensional Diversification: Improving Portfolio Selection Using Principal Component Analysis.
Managers of investment funds that invest in commodity trading advisor programs (CTAs) are faced with an abundance of investment choices. A broad array of investment styles within the CTA category and a relative dearth of transparency of the specific signals driving trading decisions amplify this. Managers of multimanager or multiasset funds generally wish to maximize returns while reducing volatility and risk of maximum loss because improving these performance characteristics tends to drive greater assets in the fund, meaning greater fee revenue for the fund manager. Investors in multi-CTA portfolios often have similar return preferences as the fund managers, and they tend to invest in funds exhibiting the best return, risk, and diversification profile. Principal component analysis (PCA) can simultaneously identify diversifiable risks and novel investment programs within a set of CTAs, hedge funds, or other investments. This capacity of PCA may be used to create more diversified CTA and hedge fund portfolios, and tests demonstrate that PCA can improve the portfolio selection process of CTAs, hedge funds, and fixed-income investments and generate ideas for implementing an alternative investment portfolio. This approach can also assist investors in multimanager funds who wish to identify the relative diversification or concentration of a multimanager fund.
Keywords: Principal component analysis, portfolio selection, diversification, commodity trading advisors (CTAs), hedge funds.