SSA_Final_Paper_1.1
ABSTRACT
Modern portfolio managers find themselves facing an increasingly challenging situation with global asset allocation. The concept of traditional asset classification has been constantly questioned yet no consensus has been reached upon in either real practice or academia. Our research attempts to answer the question of whether or not the traditional asset class definition could prove to be optimal in terms of generating the best efficient frontiers, and if it exist alternatives to reach a better solution. As most investors are comfortable with the macroeconomic logic behind the tradition classification method, few studies have been done to challenge the common-sense classification rules in the realm of asset management.
Our study begins with constructing the efficient frontiers under traditional asset allocation methodology at different levels of granularity. Theoretically, the more granular the breakdown of asset classes is, the more efficient the frontier should be. However, as the breakdown goes deeper, the adverse impact on stationarity of the portfolio as well as the statistical instability of the model causes big problems. Through out-of-sample tests on the volatility of Minimum Variance Portfolio (MVP), we prove the existence of a tradeoff between efficiency and stationarity when increasing the level of granularity in asset classes, and furthermore, find the optimal granularity of 23 asset classes (or the nearby range) for the traditional Asset Allocation Model. In addition to the traditional method, we take economic factors into the asset allocation process as a rival methodology. Unlike the traditional approach, tests on granularity through factor approach result in an increasing efficiency of frontiers and decreasing out-of-sample volatility of MVP when the breakdown of asset classes goes deeper. Last, we conduct inter-methodology comparison between out-of-sample results. The comparison concludes that the factor model outperforms the
Modern portfolio managers find themselves facing an increasingly challenging situation with global asset allocation. The concept of traditional asset classification has been constantly questioned yet no consensus has been reached upon in either real practice or academia. Our research attempts to answer the question of whether or not the traditional asset class definition could prove to be optimal in terms of generating the best efficient frontiers, and if it exist alternatives to reach a better solution. As most investors are comfortable with the macroeconomic logic behind the tradition classification method, few studies have been done to challenge the common-sense classification rules in the realm of asset management.
Our study begins with constructing the efficient frontiers under traditional asset allocation methodology at different levels of granularity. Theoretically, the more granular the breakdown of asset classes is, the more efficient the frontier should be. However, as the breakdown goes deeper, the adverse impact on stationarity of the portfolio as well as the statistical instability of the model causes big problems. Through out-of-sample tests on the volatility of Minimum Variance Portfolio (MVP), we prove the existence of a tradeoff between efficiency and stationarity when increasing the level of granularity in asset classes, and furthermore, find the optimal granularity of 23 asset classes (or the nearby range) for the traditional Asset Allocation Model. In addition to the traditional method, we take economic factors into the asset allocation process as a rival methodology. Unlike the traditional approach, tests on granularity through factor approach result in an increasing efficiency of frontiers and decreasing out-of-sample volatility of MVP when the breakdown of asset classes goes deeper. Last, we conduct inter-methodology comparison between out-of-sample results. The comparison concludes that the factor model outperforms the
Cited: 1. Bodie, Zvi, Alex Kane, and Alan Marcus. Investments. Tenth ed. New York, NY: McGraw-Hill Education, 2014. 2. Clarke, Roger G, Harindra De Silva, and Robert Murdock. "A Factor Approach to Asset Allocation." The Journal of Portfolio Management (Fall 2005) 10-21. 3. Farshid, Asl, and Etula Erkko. "Advancing Strategic Asset Allocation in a Multi-Factor World."The Journal of Portfolio Management 39, no. 1 (Fall 2012). 4. Idzorek, Thomas M., and Maciej Kowara. "Factor-Based Asset Allocation vs. Asset-Class-Based Asset Allocation." Financial Analysts Journal 69, no. 3 (May/ June 2013): 19-29. 5. Kritzman, Mark. "Asset Allocation." Windham Investment Review, November 1, 2007. 6. Kritzman, Mark, Sébastien Page, and David Turkington. "In Defense of Optimization: The Fallacy of 1/N." Financial Analysts Journal 66, no. 2 (March/ April 2010): 31-39. 7. Page, Sébastien. "The Myth of Diversification: Risk Factors vs. Asset Classes." PIMCO Viewpoints, September 1, 2010.