Despite the shortcomings of traditional asset allocation policies, most investment portfolios are still constructed based on direct asset class exposure. In addition, it may not be feasible for investors to implement policy-level decisions using a factor-based allocation framework. The authors discuss three approaches to risk factor–based portfolio construction and offer their reflections on the practical aspects of implementation.
There has been an emerging shift, especially among institutional investors, away from relatively static asset allocation based on diversification across asset classes toward more dynamic asset allocation based on diversification across risk factors. Many researchers have explored ways of incorporating risk factors into the portfolio construction process, but they tend to be primarily theoretical. The authors review three approaches to risk factor–based portfolio construction: asset class–based risk parity, alternative beta, and long–short risk premium. They discuss the investment rationale for each approach using stylized case studies and point out implementation issues that institutional investors should consider.
How Is This Research Useful to Practitioners?
The authors find that between 1995 and 2013 their hypothetical asset class–based risk parity portfolio delivered higher total returns (9.0% versus 7.2%) with lower volatility (7.9% versus 8.8%) and a smaller maximum drawdown (17.8% versus 32.4%) than a traditional blended portfolio.
Most institutional investors may find it impractical to implement asset class–based risk parity approaches. First, asset classes may not be aligned with true risk factors, such as growth and inflation. Second, investors may not be comfortable with implicit, underlying assumptions, such as the lack of a strong view on expected returns, or with inherent risks, such as interest rate risk. Third, the use of leverage may not be viable, at least outside of an alternative investments bucket.
The authors find that their hypothetical blended alternative beta portfolio delivers a higher Sharpe ratio than a traditional blended portfolio (0.8 versus 0.51). But according to the authors, institutional investors face issues and challenges in each of the five key stages of the decision-making and implementation process. Finally, the authors construct a hypothetical portfolio consisting of 10 liquid risk premiums and find that it delivers only slightly lower excess returns (4.2% versus 4.5%) with just a fraction of the volatility (2.5% versus 8.8%) and maximum drawdown (2.9% versus 32.4%) of a traditional blended portfolio.
In addition to the previously mentioned implementation issues, institutional investors also need to watch for the following when adopting a risk premium approach: (1) short selling constraints, (2) high return-eroding transaction costs, (3) unstable correlations between factors, and (4) the need for more appropriate (complex) weighting schemes than those used in the case studies. The authors conclude by emphasizing that both alternative beta and risk premium strategies require active decision making (i.e., selection of the right blend of factors and implementation strategies) on the part of in-house teams.
How Did the Authors Conduct This Research?
For their case studies, the authors use S&P Dow Jones, Barclays, and FTSE index data from December 1995 to December 2013. They construct their hypothetical asset class–based risk parity portfolio based on exposures to US equities, emerging market equities, Treasury bonds, high-yield bonds, commodities, and real estate. For the purposes of illustration, the authors opt for a naive approach, using backward-looking measures of volatility and correlations to drive quarterly adjustments to the asset class exposures.
They construct two hypothetical long-only alternative beta portfolios: (1) an equity portfolio with 40% in low-volatility equities and 60% in small-cap, value, momentum, and quality indexes; and (2) a commodity portfolio with 40% in the S&P GSCI Risk Weight Index and 60% in commodity curve, value, and momentum indexes. The authors use the alternative equity and commodity portfolios to create their hypothetical blended equity/commodity/fixed-income portfolio. They construct their hypothetical liquid risk premium portfolio by taking long positions in 10 alternative beta strategies—5 equity, 3 commodity, and 2 fixed income—and offset each of those positions against its corresponding benchmark.
The authors calculate return, volatility, and maximum drawdown metrics as well as Sharpe ratios for each hypothetical portfolio and compare them with the corresponding metrics for a traditional blended 50% equity/40% fixed-income/10% commodity portfolio. Their reflections on implementation issues are based on feedback from practitioners as well as discussions that took place in client roundtable events organized by the authors’ firm, S&P Dow Jones Indices.
The authors manage, within a relatively brief article, to provide an overview of the practical challenges faced by institutional investors seeking to incorporate risk factors into asset allocation and portfolio construction. They reference an alternative beta fixed-income component within their hypothetical blended alternative portfolio. It would have been helpful, for the sake of clarity and consistency, if they had described the construction of the alternative beta fixed-income portfolio alongside the descriptions of the alternative beta equity and commodity portfolios or at least had noted the proprietary nature of that information.