The following article was first published in the August Professional Planner magazine and can also be found on their website by clicking here … otherwise just read on …
The management guru, Peter Drucker, is attributed with the phrase, “you can’t manage what you don’t measure” and whilst we know that is not completely true, as we manage numerous tasks without measuring them every day, it is really about defining success. Having a measure of success helps define a goal to achieve and in the world of investment management, success often comes down to performance compared to a benchmark. Whilst most times this is a fairly simple, benign, and obvious issue, there are many examples where benchmarking is done poorly, can be misleading, and ultimately increase risks unbeknownst to the adviser or investor. Individual investors and customised client portfolios of financial advisors rarely have benchmarks.
This article looks at good benchmarking, bad benchmarking, plus a few tricky issues for consideration. An important part of the investment management process is performance analysis, whether up-front during the investment selection step, or during the review phase, and choosing the right benchmark should be integral.
What is a good benchmark?
In essence, a good benchmark is representative of a strategy’s investment universe and is therefore representative of its risk and return characteristics. This means some of the key characteristics of a good benchmark may include being:
- Underlying securities and their weights are clearly defined
- It is possible to passively invest in the benchmark
- Rules behind the creation of the benchmark are clear and frequently calculated
- Consistent with intended style or bias
Satisfying these characteristics is often relatively simple with little difficulty in finding suitable benchmarks for most strategies. For example, an Australian equity strategy may be small cap, large cap, or even absolute return focused, but if its mandate dictates that its investment universe is the top 300 stocks listed on the Australian Securities Exchange, an appropriate benchmark may be the S&P/ASX 300 Total Return index. This benchmark is even appropriate for the Small Cap focused strategy if it can invest in larger companies. If the Small Cap strategy is excluded from investing in the top 50 companies, then its benchmark could become the S&P/ASX 300 Total Return excluding the securities from S&P/ASX 50.
Some of the more well-known benchmarks, and their respective asset class, include:
- S&P/ASX 200 Total Return – Australian Equities
- MSCI World GR – Global Equities
- Bloomberg Ausbond Composite – Australian Bonds, and
- Bloomberg Barclays Global Aggregate – Global Bonds
Each of these satisfy the above-mentioned criteria required of a good benchmark.
Bad Benchmarks
Unfortunately, there are many strategies using inappropriate benchmarks. The main culprits are often absolute return, credit or high yield debt, and CPI-plus…and/or…pretty much any risk strategy that uses Cash (or CPI) as a benchmark. Cash (or CPI) may well be related to an investment objective, particularly given their definitions as the risk-free rate and the expectation to outperform, however, cash (or CPI) will not be representative of strategy risks.
Similarly, when a strategy invests outside of the investment universe of their benchmark, it may be time for a new benchmark. Investing outside a benchmark may change the risk and return characteristics of the strategy so comparisons can become inappropriate and therefore riskier than realized. Think bond strategies that move up the credit curve and invest in unrated securities, despite having a benchmark that may be cash or investment grade quality.
Cash will always fail the test of a good benchmark for any risky strategy as it will never be representative of the investment universe. The investment industry largely built on the belief that higher risk is required to produce higher return so over the long run, of course risky strategies should outperform cash if that risk is remotely fairly priced; but that does not mean the strategy has exhibited appropriate return or risk to achieve it.
Mixing Benchmarks with Portfolio Construction method
Many strategies will claim they are “benchmark unaware”. Being “Benchmark unaware” has little to do with measuring the success of a strategy and is more a reflection of its portfolio construction methodology. In other words, a “benchmark unaware” portfolio is probably constructed with little consideration of their asset class benchmark, may only hold “best ideas” which are weighted according to conviction of success and not their market capitalization which is a core characteristic of most liquid benchmarks. However, to ascertain whether such a strategy is successful, it is appropriate to measure against a Benchmark that is representative of the investment universe.
For example, an “absolute return” equity strategy may claim to be “benchmark unaware”, with the ability to invest in any equity market in the world. In this case, the benchmark should be more like the MSCI All Countries World Index (ACWI) and certainly not, the frequently seen, cash or cash-plus benchmark. MSCI ACWI will be far more representative of the risk the strategy and if there is likely to be significant cash holdings on a regular basis, then perhaps a strategic or expected level of cash could be included in the benchmark definition. Confusing how a portfolio is constructed does not necessarily change the risk/return profile of a strategy … as risk-adjusted excess return to a traditional benchmark still requires significant skill of a manager no matter the level of active or idiosyncratic risk.
Multi-Manager Problem … can turn into a Multi-Benchmark Problem
Most superannuation funds and financial planners, use a multi-manager approach to designing investment portfolios. Because asset allocation is a key part of the portfolio construction decision, each asset class should be appropriately benchmarked which ultimately frames the underlying manager selection towards strategies that produce the desired asset class characteristics.
Where many investors start to make mistakes (or at least increase risk), is that there is often very little consideration of the asset class benchmark, and strategy selection can become more focused on the strategy’s own benchmark. It is possible to have all underlying strategies outperform their own benchmarks but underperform the asset class benchmark.
One of the more common examples of this is the inclusion of Small Cap Australia Equity strategies as part of the Australian equities asset allocation. Numerous performance analyses over the years have demonstrated outperformance by active managers in the small cap space so the inclusion of these strategies is based on this alpha potential. The fundamental belief is that small caps are a less efficient market enabling active managers to exploit opportunities to produce excess returns. However, what is sometimes ignored is the ability of small caps to produce risk-adjusted outperformance against the asset class benchmark, which may be the S&P/ASX 300 or MSCI Australia index. Alpha amongst Small Caps does not mean Alpha amongst Large Caps.
Another example is Infrastructure. Performance analysis across the Infrastructure suite of products in Australia is often troublesome as it appears almost every strategy has a different benchmark; so understanding whether one strategy is potentially superior to another, can be difficult if looking for outperformance. Different benchmarks between strategies is simply an apples and oranges comparison.
What portfolio constructors must focus on is comparison of strategies to their own asset class benchmark and consideration as to whether a strategy will outperform it. Taking this approach should provide better insights to relative performance behavior and relative risks … apples and apples comparisons are essential for better portfolio construction decisions.
Benchmarks 2.0
With the significant growth in Exchange Traded Funds, Smart Beta, and multi-factor investing, Benchmarking is becoming a multi-layered exercise. Assessing strategies to a traditional asset class benchmark continues to be important, but the separation of determining success (or otherwise) from style or security selection is also important … you should know what you are paying for. Assessing strategies to their own style benchmark enables a deeper understanding of manager capability.
For example, it is widely accepted that a value bias across most equity markets around the world has produced outperformance compared to traditional market-cap weighted benchmarks. Largely thanks to ETFs, it is much easier and cheaper to buy style indices, like Value, so assessing an actively managed value strategy against a value benchmark will go towards understanding whether the manager is skilled at stock selection or is simply successful on the back of the systematic value tailwind. Why pay active fees, if you can get a similar result from a lower cost passive strategy that has the desired style.
What to do?
This article has touched on a few issues around benchmarking and there are many others that can be addressed another time. Either way, there are distinct lessons that can enable better strategy analysis and therefore improved portfolio management decisions. The main lessons this article hgas touched on include:
- Choose your own benchmark that reflects the investment universe of the portfolio (or asset class) you are designing
- If your investment philosophy dictates a preferred style, then choose a secondary benchmark that is reflective of that style
- Assess potential strategies against your chosen benchmark(s) to gain a better understanding of the relative risks, and of course, whether you believe there is outperformance or risk-adjusted value-add potential
- If a strategy manages to a different universe and/or a particular style than yours, assessing the manager against benchmarks that reflect their investment universe and/or style can help determine whether they are skilled, maybe lucky, or otherwise; but this is secondary to the desired characteristics of your own measure of success
Ultimately, good benchmarking is simply about creating apples and apples comparisons to better measure success. Comparisons may be return, risk, or a range of other metrics. Be careful of mixing benchmarks with objectives, do not accept benchmarks not reflective of a strategy’s investment universe; and hopefully improved measures of success will lead to more robust portfolio management and better results for investors.