Overview
Factor based asset allocation is based on the Arbitrage Pricing Theory and Factor Analysis. It decomposes asset returns into a return due to its exposure to a given risk factor (or factors) and in idiosyncratic risk. A manager may wish to use factor analysis when she has a particular view (or forecast) about the chosen risk factors. When the risk factors chosen represent a market risk, the manager may believe that the factor has an embedded risk premium attached to it. In such a case, she may wish to maximize the portfolios exposure to the various risk premia, while minimizing the portfolios risk.
Steps
- For each asset class in the portfolio, the manager identifies a set of factors that affect that asset class. Typically, managers want to have some factors that affect multiple asset classes.
- Estimate the factor exposures of each asset in the investable universe.
- Set factor targets
- Minimize the portfolio's forecasted total risk (typically standard deviation) while achieving the desired factor exposures. Note, the desired factor exposures can always be achieved as long as the manager accepts both long and short positions. In a long only portfolio, the manager may have to specify other parameters to optimize, maybe a utility function.