Op-Ed: Is Your Benchmark for Real? Probably Not

Art by Jon Han

With the close of the second quarter fast approaching, many allocators are preparing for (or bracing for!) the performance measurement process that is about to unfold once the June 30 performance reports are released. Not-for-profits with June 30 fiscal-year ending dates know all too well the heightened focus and scrutiny placed on the final performance report of the fiscal year (don’t get me started on why I believe institutions spend far too much time on peer comparisons).

Despite the uneasiness that can sometimes result from the performance evaluation process, when done properly, the exercise is one of the most important steps of the investment management process.

In this article, I suggest a small, but impactful, change to the way institutions benchmark performance. Although subtle, this small change is rooted in rationality and packs a nice punch as it produces enhancements to governance protocols (by providing a better basis to evaluate the value-add of the investment program) and improvements to manager underwriting techniques (by providing a better basis to evaluate a manager’s “alpha” potential).

Background

Throughout my investment career, I have observed the high frequency in which active managers and products fail to beat the returns of a closest fit index. However, I also observed the regularity in which returns of the index-tracking passive vehicles chosen in favor of the active manager fall short of the index return too.

In these instances, the investment team was penalized on the “performance vs. benchmark” scorecard for delivering returns below the benchmark—even in asset classes in which the implementation was passive. So naturally, questions began to swirl:

• If the index returns are unobtainable even with passive investing, why are indices being used as a basis to assess the value-add of the investment program?

• Does it make sense to continue passing on active managers that appear to underperform their benchmark only to invest in a passive vehicle that also underperforms its benchmark?

Whether the performance measurement process is used by external stakeholders to judge a team or program or by the internal investment team to help make improvements to managers, portfolio structure or decision-making processes, having an appropriate basis for the evaluation is essential. So, with June 30 just six weeks away, now seems like a good time to examine the benchmarks being used to make these assessments.

Is Your Benchmark for Real?

To consider whether or not your institution has the right basis to make clear assessments, start by asking the question: “Is our benchmark for real?” Or in other words, does the return stream being reported by your benchmark actually exist?

If the benchmark you are using is comprised of commonly used indices, then the answer to this question is unequivocally “No!” Your institution cannot buy or own the return being reported by your benchmark.

If an institution cannot own the return stream of its benchmark, does that mean the returns being reported by index providers are fake news? Of course not! Index providers play the very important role of selecting and weighting a basket of investments that, in their view, best represents a given segment of the capital markets. We need the work of these index providers to help us identify and quantify the various opportunity sets available.

However, we must use the reported index data properly. Index data can and should be used to help investors set asset allocation targets and to construct portfolios; but index returns should not be used as a basis to quantify the value-add of an investment program (or a manager). Why? Well, simply put, indices are not investable.

But What About Passive Vehicles?

Yes, there are many products designed to track a given index. However, as the analysis at the end of this article shows, these products do not capture the exact returns reported by the indices. Index tracking vehicles have embedded costs and their returns are affected by the cash flows related to investor activity. Therefore, these vehicles will not be able replicate the return streams of the cost-free and unencumbered indices. Sometimes the difference in results is small, other times it can be quite substantial.

But either way, what is the better basis to judge the value-add of the investment program—versus the unobtainable return of the index or versus the real return of the investable passive vehicle?

The CFA Institute provides additional perspective on this topic. The Institute posits that for a benchmark to be a valid and effective tool for measuring performance, it must be:

• Unambiguous

• Investable

• Measurable

• Appropriate

• Reflective of current investment opinions

• Specified in advance

The commonly used indices meet all of these criteria … except one. The indices are not investable.

So, if we seek to align our performance measurement processes with the CFA Institute’s suggested protocols, what can we do? How do we solve for the shortfall stemming from the fact that the commonly used indices are not investable?

Solution

The solution is actually straightforward. Each institution should create its own unique passive implementation portfolio. To create this portfolio, the institution would identify the specific mix of products/managers/vehicles it would use to passively execute its policy portfolio. The institution will factor in its own unique characteristics, such as position size, liquidity needs, cost sensitivities, etc., to determine its best fit passive vehicle for each asset class represented in its policy portfolio.

These selected passive vehicles, mixed together in the weightings of the target portfolio, create an investable policy portfolio. Because each institution will account for its own unique factors when choosing the best fit passive vehicle, no two institution’s passive portfolios will be the same. Once constructed, this passive implementation portfolio has real returns (its returns are achievable), and therefore, the returns of this passive portfolio can now be used to evaluate the value add of the investment program (see table).

Improving the Basis for Evaluating Value-add

Investment committees, boards, and other stakeholders often try to quantify the impact of the investment decisions made by the investment team/program. Because index-based benchmarks report returns that are not investable (not achievable), their use as a basis to judge the contributions/detractions of investment decisions is ineffective.

However, switching from an index-based policy benchmark to an investable policy benchmark (investable passive portfolio) can solve for this. The investable policy portfolio tells the IC/Board exactly what the institution’s return would have been through passive implementation of the policy targets. This return can be used to compare the return generated by the team/program to help quantify the impact of investment decisions made by the team/program. A real versus real comparison!

The investable benchmark is also a better basis upon which to structure an excess return-based incentive plan. Compensating the team for excess returns above what the institution would have earned through passive implementation of its policy targets very effectively aligns interest while also removing the noise associated with comparing returns to the uninvestable, index-based benchmarks.  

Because the investable policy portfolio (the passive portfolio) will typically have a lower return than the index-based return, the switch might conjure up the notion that, “by switching to investable benchmarks, all you are doing is lowering the bar.” No high-quality investment office or investment committee should ever be comfortable with a measurement process that simply lowers the bar.

In my opinion, the suggested solution in no way lowers the bar. In fact, some of the selected index-tracking vehicles actually outperform the reported index returns for certain time periods (securities lending and favorable effects from cash flows can explain this, with the former being an important incremental risk factor that should be carefully scrutinized prior to investing).

More pointedly, the creation of investable policy portfolio produces a real return—or a return that is actually achievable by the institution. This achievable return factors in the realities of implementation (costs, liquidity, access, etc.) and therefore provides a much better basis to judge the effectiveness of the investment program.

And even if you are not comfortable making a formal change to the policy benchmark, still consider going through the process of creating a passive, investable portfolio. The exercise will be illuminating and may even spark some useful and much needed resourcing and prioritization conversations. Further, consider adding the returns of this passive, investable portfolio to your performance reports. Its inclusion on the reports will give rise to constructive conversations around active manager choices.

Plus, knowing the passive alternative for each asset class in your portfolio can enhance the manager selection process, as described below.  

Improved Manager Selection Decisions

By using an investable passive vehicle (attainable) return rather than an index (unobtainable) return as the basis of comparison for an active manager, the investment team can garner a clearer assessment of a manager’s potential return contribution.

In my opinion, it is more fitting to assess a manager’s possible alpha by determining the manager’s potential excess return above the investable passive alternative rather than relative to the asset class index (which is not an investable return). Although it may seem a bit odd, some active managers may show long-term past performance below a given index, but above the index-tracking (passive) vehicle. High yield may be a place this could occur given the wide variance in performance between the index and the available passive vehicles.

Isn’t effective diligence supposed to help us determine the better investment? If so, and if you are trying to determine where to direct dollars, then comparing two real choices is logical. Switching the underwriting process of an active manager away from one in which the manager is compared to the unobtainable return of an index to the real return of a passive alternative will lead to return enhancing outcomes over time.

Investable Benchmarks for Alternative Investments

While the investable benchmark process illustrated at the end of this article focuses on the long-only part of a portfolio, there are ways to develop investable benchmarks for hedge funds and private investments. A full explanation of how this works would require adding substantial length to this article (congrats if you made it this far), so I will spare the additional detail, at least for now. In short, as with traditional benchmarking for alternative investments, the investable benchmark approach has limitations; however, the investable benchmark approach produces a much better basis to judge value-add than the typical alternative investment benchmarking process being utilized by most institutions today.

Conclusion

If you are an investment committee member in search of processes that inspire investment excellence, encourage the institutions where you serve to adopt an assessment process that more clearly quantifies the value-add of the investment office (comparing portfolio returns to real available returns).

If you are an allocator, encourage your governing body to use the passive portfolio as a basis to compare the impact of your investment decisions. Even if your institution does not officially change its benchmark to an investable benchmark (as I highly encourage), add a line to your performance reports showing the return of a passively invested policy portfolio. This return will be lower than the index-based return, so explain that you are not trying to lower the bar, but rather simply aiming to illuminate the real passive return available to the institution. Also, shift your active manager underwriting process to one that compares manager returns to an investable passive vehicle rather than the unavailable index return. This switch in techniques will help you make return enhancing decisions over time.

If you worried about living in a world dominated by fake news, shifting to an investable policy benchmark will help make things more real. The benefits of comparing real (achievable) returns to actual portfolio returns can lead to better assessments (enhanced governance protocols) and produce better manager selection decisions (better performance).

A subtle, small change that packs a nice punch.

Investable Index Explained

If you are interested in an example of how to create an investable index, see below.

Take, for instance, a simple $2 billion portfolio in which the long-only portfolio represents half of the portfolio (or $1 billion) that maintains the following policy target weights:

The Policy Benchmarks might be represented by the indices specified below:

It would be common for the Policy Benchmark for this portion of the portfolio to be a roll-up of the asset class indices, weighted in accordance with the policy targets. Using that methodology, the average annualized five-year return ending March 31, 2020, for the long-only policy benchmark equals 2.59%. According to the indices, the long-only portfolio would have grown from $1 billion to $1.136 billion during this time.

However, was this 2.59% return available to the institution? What would have been the actual return of the long-only portfolio had the institution wished to execute this policy portfolio passively?

Factoring in its size, liquidity needs, and cost sensitivities, the institution may have determined that the best-fit passive vehicle for each asset class to be the institutional mutual funds listed below. If so, this passive portfolio would have achieved a return of 2.51%, or 8 basis points (bps) less than reported. Passive implementation of the long-only portfolio would have grown this part of the portfolio from $1.000 billion to $1.132 billion, or $4 million less than reported by the indices.

Not bad. But try to explain to your stakeholders that the 8 bps, and $4 million of gains, reported by the benchmark were truly not available.

If a particular institution desires the intra-day liquidity provided by ETFs, the passive implementation of this policy portfolio would have returned 2.36%, or 23 bps less than reported by the indices. And the long-only portfolio would have grown to $1.123 billion, or $13 million less than that indicated by the indices.

Think about the time spent in a committee meeting explaining the shortfall in returns simply because the indices produce returns that are not obtainable. Wasted time! This noise can be eliminated with the use of investable benchmarks.

Maybe these return differences do not sound meaningful, but think about trying to make up for what appears to be a $4 million to $13 million shortfall of available returns. Of course, because the index returns are not investable, there is no actual shortfall—only the appearance of a shortfall. In reality, the investor captured the highest passive return available to the institution.

If this seems like small dollars, think about the effects had a longer time period been used in this example. Or your own institution’s impact if your portfolio’s long-only portfolio is larger than the $1 billion used in this example. And notice from the data below that a greater weighting to some of the higher tracking error passive vehicles (high yield, emerging market, etc.), will widen the gap between the reported index return and the passive portfolio return. Moreover, this illustration does not factor in a more complex portfolio structure. If your benchmark includes representation of indices in less used asset classes (converts, non-US small-cap, etc.), the difference between reported returns and achievable returns will likely be even greater.

For reference, here is a table reflecting the difference between the index return and the available passive return. The passive vehicles were chosen simply as examples only. Your institution would factor in its own size, cost sensitivities, liquidity needs, etc. to select its own preferred passive vehicle. That chosen vehicle will be used to construct your institution’s investable passive portfolio.

Good luck! The time and attention to this exercise will be worth it.

*Key

Source: Neuberger Berman, Vanguard, SSGA, Blackrock, Fidelity

Past performance is no guarantee of future results. Performance shown is net of fees.

¹Returns before taxes
²Total return represents changes to the NAV and accounts for distributions from the fund.

This feature is to provide general information only, does not constitute legal or tax advice, and cannot be used or substituted for legal or tax advice. Any opinions of the author do not necessarily reflect the stance of Institutional Shareholder Services or its affiliates.

Related Stories: 

Op-Ed: A Tough Time for the Credit Markets

OCIO Roundtable (Part 1): Portfolio Rebalancing During COVID-19 Mayhem

Op-Ed: Rethinking Your Inflation-Protection Strategy