The choice of benchmarks is one of the most fundamental decisions that investors make, and yet it gets little time or respect in the investment process. I want to explain why benchmark selection matters, and why investors should care about it.
We should start with the most basic idea. All benchmarks are simply portfolios. Like any other portfolio, the assets held, and the weights of those assets, is what drives behaviour. Benchmark portfolios become important only because we believe they have special characteristics: that they act as good representations of a particular market or sub-market. Of course, it isn’t controversial to say that the risk and return characteristics of two portfolios will differ. It shouldn’t therefore be controversial to say that different benchmarks which are trying to describe the same asset class can have significantly different risks, returns, and factor exposures.
What drives these differences? Simply the portfolio construction methodology, as you would expect with any set of portfolios. Different index providers calculate their indexes in different ways, with different approaches. For example, indexes may reflect free float in different ways, which impacts weighting structures accordingly. Each index is designed for a particular purpose and with a different philosophy: each potentially valid, and each with strengths and weaknesses.
The interesting thing about benchmark selection is the degree to which it impacts every other investment decision you make as an investor. All of your decisions will be assessed relative to the benchmarks you have picked to represent each asset class. All of the sophisticated risk analysis that you perform will be done relative to the benchmarks you have picked to represent each asset class. Your very success and failure as an investor will be driven by those decisions as well. It can be amusing to compare the degree of precision with which we do so many of the calculations we use in the investment process, while spending little time examining the underlying benchmark we are using as the measuring stick for those highly precise calculations.
One way that this can directly impact the portfolio is in the assessment, measurement, prediction and allocation of tracking error. Investors want to make sure that they only incur tracking error in places where they believe that tracking error will be compensated. This requires, of course, an accurate measurement of that tracking error.
A second important impact can be in the assessment of value created by managers. There is an incentive in place for managers to try to “game” the benchmark against which they are measured. This is particularly obvious in the fixed income space: a manager who is building a portfolio of riskier credit assets but who is benchmarked against a core type benchmark with safer government assets will be able to generate long-term outperformance simply because of the underlying mismatch of those portfolio structures. This might be attributed to manager skill, when in reality some or all of it is perhaps better classified as benchmark mismatch.
A similar effect can be seen in international equity portfolios. Managers assessed against a benchmark which includes no emerging market exposure may decide to allocate assets to emerging market securities. Since over the long term these securities are expected to provide a higher return, this may benefit portfolio performance, assuming the manager selects the securities effectively. The challenge, of course, is that the owner of capital may incidentally over-allocate to the emerging market asset class unless they understand these actions of their international equity managers. The asset owner may also unfairly view the international equity manager as skilled, when in fact the manager’s performance may be better attributed to an asset allocation decision. In each case the skill and effectiveness of the manager are mis-represented: better benchmark selection would avoid this problem.
A third problem can be particularly important at large organisations with significant investment staff directly employed by the plan. One of the challenges owners of capital sometimes face is being able to demonstrate the value created by highly talented internal investment staff, so that market level compensation can be awarded to those people. This requires careful selection of benchmarks to properly assess each part of the portfolio that these staff are directly responsible for. As these benchmark selection decisions are at the heart of the employment relationship, and direct staff compensation is dependent on accurate benchmark selection, these decisions become truly core to the mission of the organisation.
So what can investors do? The most important thing is simply ask questions about benchmarks. Consultants and index providers should have strong capabilities in this space, and should be able to provide detailed advice to investors on the best benchmark selection policies to adopt. These should be revisited regularly, to make sure that the benchmarks that are being used at the asset allocation and portfolio construction level are sensible, effective, and provide good insight into the portfolio structure that has emerged from those allocations.
Tools are available to help them make this decision, and detailed risk analytics systems can be used to help understand the underlying risk structures embedded in each benchmark. It is not incoherent to compare benchmarks to one another as though they were each portfolios – because that is exactly what they are. Spending a small amount of time validating these important choices can help tighten up each step of the process thereafter. The more effective the benchmark choice, the more likely all of the subsequent analytics are to be accurate. Accurate analytics don’t guarantee exceptional results of course, but they certainly represent a good start.
Ian Toner is CIO at Verus Investments.