From simple beginnings, the investment benchmark has become a de-facto aspect of most investment programs today. There are many types of benchmarks, but the most widely accepted version uses passive market indices as standards of performance measurement and points of comparison for various portfolio characteristics. The ubiquitous nature of benchmarks has led index providers to create passive indices representing just about every asset class, country, and investment style in the market today. But while this has made access to data much easier, it has also created a whole new set of issues for investors: with all this complexity how does one go about choosing a benchmark that’s appropriate for them? This paper will explore the methods that investors can use to build a customized benchmark. In addition, it will explain how the results of a performance comparison will depend on the method of benchmark construction. As a quick reference guide, the appendix of this paper contains definitions for many common indices used in benchmark construction today.
Finally, investors should keep in mind that, despite their widespread use for relative performance evaluation, benchmarks based on passive market indices still can’t fully evaluate whether an investment program will achieve its ultimate goal: meeting investors’ objectives. After all, a portfolio’s relative performance could exceed its benchmark every year, yet its absolute performance could still fall short of what is needed.
The Benchmark Construction Process
Broadly speaking, Investors can organize benchmark construction into three basic steps:
Step 1: Understand the portfolio and the investment manager. This is important because it lays the foundation for the other steps and the benchmark construction process in general. Important questions to gain an understanding around include:
- What are the portfolio guidelines, including permissible asset classes for investment and allowable asset allocation ranges?
- Are there any outright restrictions on types of securities or transactions?
- How does the portfolio manager seek to add value relative to their universe? The two main ways to define this are (1) security selection and (2) asset allocation. Security selection means a manager adds value by selecting specific securities that outperform the general asset class in which it’s grouped. In asset allocation, value is added by overweighting outperforming asset classes and/or underweighting underperforming asset classes. Of course, managers must have the ability to allocate across multiple asset classes for this to happen, and for those that do, their performance relative to a benchmark is often the combination of security selection and asset allocation decisions. The tables below show examples of these effects working to add value.
Step 2: Select a passive index or indices to represent the portfolio’s investible universe. This step requires in-depth knowledge on the composition of passive indices and how well the indices represent certain asset classes. Suffice it to say that investors should seek to select passive indices that best represent a portfolio’s universe as defined in step 1.
Step 3: Develop rules to calculate the actual benchmark while ensuring it remains neutral to any active viewpoint that a portfolio manager might have. “Rules” in this context refers to the process taken to actually calculate the performance of a benchmark. The process is obvious for a single asset class portfolio (i.e., simply take the performance of the representative index). However, for multiple asset class benchmarks, performance and characteristics must be weighted and combined in some way. A weighted average, as illustrated below, is the most widely accepted approach.
Applying the Process to a Portfolio
After applying these steps to actual portfolios, what does an effective benchmark look like and how can an investor interpret it? This section will apply the guidelines above to two hypothetical portfolios, one managed to a single asset class and another managed to multiple asset classes.
Single Asset Class Portfolios
Step 1: Understand the portfolio and the investment manager. Examples of single asset class portfolios might be labeled as “U.S. Large Cap Core” or “Corporate Fixed Income”. As their names imply, the mandates of these portfolios prevent them from deviating too much from one asset class based on domicile, style or sector. A typical restriction in this type of portfolio might say something about minimum permissible market value, avoiding foreign domiciled companies, a minimum bond credit rating, etc. Because they operate primarily in one asset class, portfolio managers can add value in one major way: through security selection (i.e., selecting securities that outperform the general asset class in which those securities are grouped). A single passive index, representing the universe of securities that the manager is permitted to invest in, is generally the best way to benchmark these types of portfolios and judge the manager’s skill in selecting securities.
Step 2: Select a passive index or indices to represent the portfolio’s investible universe. The best way to match up a passive index with a single asset class portfolio is to compare the characteristics of the index with the characteristics of the portfolio’s investible universe. Many managers already utilize passive indices as a starting point for portfolio construction so in this case it’s just a matter of using that same index. If the manager’s process is more nuanced, it might be advisable to take account of the restrictions placed on the portfolio as uncovered in Step 1, and then find a passive index or indices that approximate those restrictions. It should be noted that there is rarely one index that can perfectly capture all of a manager’s investible opportunities. However, using a broad representation to approximate the universe can still provide for meaningful comparisons.
Step 3: Develop rules to calculate the actual benchmark while ensuring it remains neutral to any active viewpoint that a portfolio manager might have. Since the benchmark for a single asset class portfolio is usually represented by a single passive index, deriving the performance of the benchmark is straightforward: simply use the performance of the passive index. This will also reflect the requirement that the benchmark be neutral to active viewpoints because the benchmark simply weights each security in the index in a passive way.
Multiple Asset Class Portfolios
Step 1: Understand the portfolio and the investment manager. Multiple asset class portfolios are characterized by their ability to invest in many different asset classes and use many types of securities. Sometimes, the managers of these portfolios will have wide flexibility in what to invest in and will vary their allocations in an attempt to take advantage of opportunities in the market. Other managers may have little or no flexibility in this way. Their portfolio perhaps more closely resembles a combination of single asset class portfolios. If a portfolio does have the flexibility, then it can add value in both of the ways already discussed: through asset allocation and through security selection. It will be important to understand if allocation ranges are permitted and if there are any meaningful restrictions on types of securities. For both flexible and static approaches to allocation, a benchmark is best represented by combining indices into one track record, but the method of that combination and what the benchmark tells you will differ depending on the flexibility of the portfolio.
Step 2: Select a passive index or indices to represent the portfolio’s investible universe. Similar to the single asset class portfolio, the characteristic of each asset class within a multiple asset class portfolio should be compared to the characteristics of appropriate passive indices.
Step 3: Develop rules to calculate the actual benchmark while ensuring it remains neutral to any active viewpoint that a portfolio manager might have.
If multiple indices are being combined to form the benchmark, what weight should be allocated to each index? As mentioned on the previous page, multiple asset class portfolios have an additional degree of complexity because often, the portfolio manager has the flexibility to allocate actively between different asset classes. Remembering that the benchmark is supposed to represent the portfolio but be neutral to any active decisions, multiple asset class benchmarks can remain neutral by using expected long-term allocation averages for asset class weights. This could effectively be represented by a portfolio’s long-term strategic targets or the midpoint of allowable ranges. For example, if a portfolio’s stock exposure can range from 40% to 60% of assets, and its long-term target allocation is the midpoint of that range (i.e., 50%), then the weight applied to the stock index in the blended benchmark would be 50%. This method also provides investors with a more complete way to assess a portfolio manager’s success in adding value because the benchmark is neutral to both asset allocation and security selection decisions.
What sort of rebalancing method should be used for multiple asset class benchmarks? “Rebalancing” in this context means the frequency with which the weighted average calculation (depicted on page 3) is applied across multiple indices. Common frequencies include a monthly or quarterly time frame, but investors should realize that if a benchmark rebalancing gives a larger weight to an underperforming index, overall performance will be lower, and vice versa. Because of this, different rebalancing frequencies will show different performance results even though the weights and indices are the same. The example below shows actual historical performance by decade for a hypothetical 50% stock, 50% bond portfolio. It also shows how much performance outcomes differ depending on the rebalancing frequency. Across the four methodologies, performance in some decades can be quite different or similar. The magnitude of the difference should generally depend on the volatility of performance over the time frame.
A fair and appropriate benchmark is an important tool in assessing a portfolio manager’s investment skills. The benchmark construction approach, as described on the previous pages, should let investors build a benchmark that is representative of their portfolio and accounts for the different ways in which a manager can add relative value, whether that manager uses a single asset class or multiple asset class approach.
With that being said, it’s also important to remember the limitations of using relative benchmarks. For example, context in terms of the market environment or the time period over which an evaluation is taking place is important to consider. It’s also important to have a deep understanding of an investment manager’s process and the expectation for when and how they expect to add value.
Finally, it’s also important to note that most portfolios aren’t just designed to meet relative performance targets; they’re ultimately designed to meet some future liability for the investor. For example, individuals in retirement and corporate pension plans both have absolute oriented goals with regards to meeting a future spending level. An investment benchmark might be able to provide relative context for portfolio evaluation, but no evaluation would be complete without delving into what the portfolio is doing to manage risk, meet its liabilities, and ultimately, help investors achieve their goals. We believe both types of evaluation are needed to get a complete picture of a portfolio.