Long-term investors will need to adopt next-generation strategic asset allocation practices to successfully play their part in a rapidly changing world. Truths, practices, and principles traditionally accepted by long-term investors face reduced significance and potential obsolescence. We will examine the key topics and expected changes to how long-term investors will make investment decisions in a series of articles on multi-horizon strategic asset allocation applications. In this article, we examine the benefits of dynamic rebalancing vs. the constant-mix (fixed-mix) approach in strategic asset allocation.
Multi-Horizon Analysis Recap
In our previous article, we introduced multi-horizon analysis and argued that few long-term investors take a purely buy-and-hold approach to their investments. Constant-mix (fixed-mix) was presented as a practical alternative, whereby an investor rebalances back to the original allocations periodically. The impact of doing so changes the investment risk/return and presents a different efficient frontier to the investor.
A single set of weights for the entire investment horizon ignores many investors’ flexibility in changing their asset allocation to exploit investment opportunities and manage risk.
Why Think About Dynamic Asset Allocation?
Few investors expect that asset class return and risk will remain static throughout the entire investment horizon. Allowing weights to dynamically change over time to take advantage of market opportunities is a natural extension to the fixed-mix approach. We refer to this as a dynamic-rebalancing strategy or dynamic asset allocation.
To illustrate, we use the following Capital Market Assumptions (CMA) inputs. Risk (volatility) is estimated from historical returns data.
Table 1: Forecasted Returns and Volatility Across Asset Classes
Forecasted Return (Annualized)
When using a traditional single-period asset allocation approach, an investor with a 20-year investment horizon is forced to select a single set of return forecasts when constructing the optimal allocation. The appropriate choice is not obvious as there are several available options:
Use the return forecast that matches the investment horizon (20 years). In this case, the portfolio is likely to overexpose the investor to weak fixed-income returns for the first five years.
Use the return forecast in the immediate period (five years). One could argue that since you only set your portfolio allocations for the immediate future, only the shortest-term forecast matters. Furthermore, the shortest-term forecasts are likely to have a smaller error associated with them. Taking this approach effectively ignores the longer term, and successive short-term allocations could lead to a long-term average allocation quite different from one constructed with the long-term inputs.
Use an average of the return forecasts. Using averages leaves out opportunities to be exploited in the short and long term. This is analogous to monitoring only duration, which omits important nuances in a barbell fixed-income strategy.
Dynamic asset allocation doesn’t force a choice and allows you to use all the information in Table 1. At each period, the portfolio is optimized based on all information provided; weights are allowed to dynamically change period-to-period to exploit investment opportunities. Focusing on the Credit 3-10-year asset class, we can see that it is expected to experience the most significant improvement in returns over the investment horizon. Intuitively, we’d expect to see a corresponding increase in allocation throughout the investment horizon.
What Does Dynamic Asset Allocation Look Like?
Below is the ex-ante evolution of portfolio weights for a dynamically rebalanced portfolio compared to a fixed-mix portfolio with the same inputs from Table 1.
Why Use Dynamic Asset Allocation?
Dynamic allocation provides a much richer output and drives healthy analysis, observations, and discussions between investors and investment managers. Enhancing the communication and understanding of the investment strategy and likely outcomes helps to promote long-term investment behaviors in managers and investors alike.
Examples of areas enhanced by dynamic asset allocation are:
Validation of expectations. The increase in allocation to Credit3-10 year from year 10 onwards corresponds with our observation of the increased expected return of the asset class.
Communication tool. It’s a lot easier to explain to an investor why they should care about the LongGovt and Credit3-10year asset classes even if the allocation is initially low if they can see the future importance of the asset classes.
Challenging of input modeling/assumptions. The detail in the output can expose deficiencies in input modeling and assumptions that would otherwise go unnoticed. For example, LongGovt and Credit3-10year trade places in the portfolio over time. Is that reasonable and expected?
Ability to overlay implementation considerations. A turnover constraint of 40% introduces some "smoothing" of the allocation changes. E.g., aggregate bonds take three years to drop from the portfolio; the turnover budget was better used elsewhere during the initial years. Similarly, it takes several years for the allocation to Credit3-10year to ramp up fully.
Realistic reflection of ramp-up and lock-up characteristics of private investments. Allocation to private asset classes can be allowed to gradually increase during ramp-up. Then, they can be fixed for the lock-up period and allowed to change dynamically once the lock-up period ends.
More sophisticated rebalancing decision rules (management rules). Management rules are triggered by reaching certain thresholds such as Constant Proportion Portfolio Insurance (CPPI) or minimizing cash contributions to the portfolio.
Not surprisingly, allowing portfolio weights to change over time results in a greater range of risk/return possibilities. Critically, where the range coincides with other strategies such as buy-and-hold and constant-mix, the dynamic strategy's efficient frontier dominates other strategies.
Allowing for dynamic rebalancing is a powerful component in the next generation of strategic asset allocation tools. Providing an idea of how weights are likely to evolve over time allows for better questions to be asked and a greater understanding of the characteristics of a particular investment strategy.
Our next article will explore how multi-horizon analysis can help provide more suitable analytics and approach for long-term investors to manage risk.
Todor Bilarev, PhD, Senior Quantitative Researcher, Analytics and Trading Solutions, contributed to this article.
Disclaimer: This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.
Principal Product Manager, Analytics and Trading Solutions
Mr. Chen Sui is a Principal Product Manager for Analytics and Trading Solutions at FactSet. In this role, he is focuses on helping clients adopt long-term investment behavior through the adoption of ideas and practices best suited to long-term investors. He has over 15 years of experience working with analytics clients across the globe and a wide variety of buy-side institutions such as traditional institutional asset managers, asset owners, hedge funds, and wealth management firms. Prior, he spent time in performance, risk, project, and client-servicing teams at Wilshire Associates, JP Morgan, and CQS. Mr. Sui earned a master’s degree in Mathematical Finance from the University of York and is a CFA charterholder.
The information contained in this article is not investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.