flexible indeterminate factor-based asset allocation 201606 jpm
This journal article (The Journal of Portfolio Management, 2016, Vol. 42 No. 5, pp. 79–93) introduces Flexible Indeterminate Factor-Based Asset Allocation (FIFAA), a four-step hybrid framework that combines modern portfolio theory, factor investing, reference portfolios, and bottom-up “best ideas.” FIFAA separates strategic portfolio choice into a smaller set of systematic factor exposures (the stable core) and a larger, indeterminate set of asset-class implementations, enabling institutions to explicitly target factor exposures while preserving flexibility to incorporate less-quantifiable preferences. The paper discusses factor selection, empirical mapping of asset classes to factors (authors favor lasso with cross-validation), methods to resolve indeterminacy (pseudo-inverse/inversion, candidate-portfolio filtering, and a second-stage optimization), rebalancing ranges via mapping perturbations, limitations, and avenues for further research.
Questions this document answers
The Journal of Portfolio Management, 2016, Vol. 42, No. 5, pp. 79–93.
Stephen Blyth (President & CEO, Harvard Management Company; professor of the practice of statistics, Harvard University), Mark C. Szigety (senior vice president at Harvard Management Company; managing director of research at 3EDGE Asset Management), and Jake Xia (chief risk officer, Harvard Management Company).
Four steps: (1) selecting factors, (2) measuring asset class/investment universe factor exposures, (3) choosing desirable factor exposures, (4) determining asset-class targets and ranges to achieve those factor exposures.
A parsimonious set of between four and six factors.
World equities; U.S. Treasuries; high-yield credit; inflation protection; currency.