Axioma introduced the Alpha Factor in March 2006. The following questions and answers provide insights on the Axioma Alpha Factor and its features.
What is the Axioma Alpha Factor?
Axioma Alpha Factor is a patent-pending method that dramatically improves risk forecasts of factor models. The Alpha Factor is built into Axioma Portfolio, making it the most powerful and intuitive portfolio risk analysis and rebalancing tool on the market today.
How does it work?
Portfolio managers often find that their portfolios have a level of realized risk, or realized active risk, that is different from their target, or predicted, values. Over- or under-prediction of the volatility of a portfolio is often referred to as the bias of the factor risk model.
Axioma's method uses the portfolio to be analyzed to deduce one or more "missing" factors that are orthogonal to the original factors in the risk model. The presence of the additional portfolio factor recovers a component of risk that otherwise remains unaccounted for. This adjustment efficiently mitigates the impact of the most vexing and stubborn problems in investment management today: model risk, estimation risk, and underestimation bias inherent in portfolio construction.
How does Axioma's solution differ from other attempts to build "hybrid" statistical/fundamental factor models whose statistical factors are also orthogonal to fundamental factors?
Other hybrid approaches find principle components that model risk residual to the original fundamental factors. The computations are based on statistical analyses over large populations of securities in a market. The model provider locks in a set of principle components based on their explanatory power (high Eigen values), then computes each security's exposure to those selected principle components. The presumption is that those unnamed factors are universal, pre-specified, independent of the original universal factors, and clearly independent of any individual portfolio.
A key aspect of the Axioma Portfolio Alpha Factor is that the adjustments to the original factor model (the additional factor covariances and security exposures) are not known in advance and cannot be determined until a specific portfolio has been selected and analyzed. By tailoring the portfolio factor to the individual portfolio, Axioma identifies the most parsimonious and powerful compensation for estimation error, model error, and risk underestimation bias. Note that in the limit (when analyzing an entire market as the "portfolio" in question), other hybrid approaches may converge on Axioma's portfolio factor. This is a special case of Axioma's more general solution. But for practical, institutional portfolios (diversified across perhaps 50 to 200 securities), the complexion of the specific portfolio will result in a unique portfolio factor (from among the huge family - on the order of 10,000 in the US equity markets alone-of possible orthogonal factors) that best re-captures the risk "missing" from the original structural factors.
