In multi-factor equity index construction, the decision-making and practical implementation can be complex and challenging. This paper examines the range of portfolio construction choices available to those seeking rank-based, multi-factor approaches, and the relative advantages of each.
Through back-testing hypothetical portfolios based on the S&P 500®, this paper evaluates the following construction choices: top-down versus bottom-up; sector-neutral versus sector-agnostic; portfolio concentration; weighting scheme; and rebalancing frequency. To measure the effectiveness of each portfolio, a factor efficiency ratio (FER) is proposed, which allows investors to gauge their factor purity without having to invoke the complexity of a risk model.
Our paper concludes with key findings, including the following.
- Sector-neutral portfolios may be more efficient than sector-agnostic.
- Top-down approaches may dilute exposures but are still efficient.
- Factor score-based weighting schemes may improve efficiency.
INTRODUCTION
The benefits of diversifying across a multitude of smart beta equity factors have been supported and explained in a host of research and literature. Single-factor indices (quality, value, momentum, low volatility, and small size) may reward market participants over the long term, but can be notoriously difficult to time over the shorter term. Multi-factor indices, on the other hand, generally forgo the need to time each factor and instead, through deliberate diversification, may provide more stable excess return outcomes.
Multi-factor equity investing may be well justified in theory; however, there are numerous practical portfolio construction choices to consider, each with its own advantages and implications. Without a consensus on the most effective multi-factor technique, indices offered in the market have fractured into a variety of vastly different methodologies. Some employ optimization and risk models to determine the most effective portfolio based on the strategy’s objectives. Others dismiss the complexity and lack of transparency of optimized solutions, instead favoring the relative simplicity of rank-based selection rules. Yet even within this latter realm, the choices may appear countless and overwhelming.
In this paper, we attempt to demystify the range of choices available to market participants seeking rank-based, multi-factor approaches. In doing so, we compare the relative advantages of each approach and discover the trade-offs between each decision. Critically, the approaches should be measured both on their effectiveness and efficiency in terms of risk.
Importantly, we do not advise investors which strategic factor allocation decisions are the more successful. Also, testing the robustness of multifactor performance across markets and time periods was outside the scope of this paper.
Instead, by testing only one multi-factor combination on the S&P 500, the paper’s purpose is only to demonstrate relationships among portfolio construction choices. Our goal is to arm market participants with the necessary knowledge to help determine which multi-factor portfolio construction techniques are most appropriate for their own investment objectives.
