Benchmark Misfit Risk: Is Your Fund Selection Process Undermining your Asset Allocation?

OVERVIEW OF THE PROBLEM

Most portfolio managers believe that the set of funds they select delivers the market and factor exposures of their asset allocation, as reflected in their multi-asset class benchmark. Sadly, nothing could be further from the truth. Lurking below the surface in their performance reports is a substantial mismatch between their expectations and the reality of what their funds deliver in terms of return and risk.

The cause of this decoupling? It's the "alpha seeking" efforts by the managers of each of the underlying funds, whose active decisions produce individual funds that don't really act like their promised mandates. As a result, a surprisingly high percentage of the portfolio's active risk (i.e. "tracking error") is the result of what we call "Benchmark Misfit Risk."

This is the risk that the portfolio's return pattern aligns more closely with a different asset allocation than the stated benchmark, so that much of the portfolio's active risk that we assume is driven by investment selection is really the result of investing in "out of mandate" securities. Is this a "big deal?" In a word... YES.

In this article, we will show how to identify this problem, what to do to reduce this unintended and unmanaged risk, and how to use this approach to build better portfolios that deliver a much higher likelihood of delivering true alpha instead of the "bait & switch" situation that clients potentially face.

This article was initially published in the Fall 2022 edition of the Journal of Performance Measurement, along with an abbreviated version in the CFA Institute's Enterprising Investor blog:

The original material for this article made its debut in presentations over three days in 2022 to the CFA Societies of Sweden, Finland and Denmark, and was presented most recently at the 2023 PMAR conference sponsored by TSG Performance.

IT STARTS WITH AN ASSET ALLOCATION

Here is a fairly typical "70/30" strategy of global stocks and bonds, with finer allocations to the segments within each asset class. This also reflects the mix of the performance benchmark.

Here's another view of the strategy, a bit easier to visualize in terms of relative asset positions.

WHAT'S IN A NAME?

At this point, we must ask this critical question:

"Do you want a portfolio that LOOKS like the benchmark, or ACTS like the benchmark?"

Using a decades-old technique typically called "style analysis" we examined the portfolio's return pattern to solve for the set of market exposures that produced the closest match. This set of "effective exposures" shows what the portfolio "acts like" and indicates the collective market exposures across all of the portfolio's funds. The difference from the claimed exposures is quite surprising...

The portfolio's effective exposures reflect a more aggressive set of "growth-oriented" market exposures. This is reflected in the following 5-year performance results:

Not surprisingly, these uncoordinated and unmanaged cumulative deviations from the target allocations produce a strategy that is somewhat less efficient, although thankfully it is not a substantial source of total return inefficiency.

HOW DID WE GET HERE?

By examining the "effective exposures" of each fund, we find that on average the funds hold investments that "act like" their narrow mandate about half the time. However, the deviations are often into sectors that are strongly-related to the target allocation. For example, the US Large Cap Growth fund held over half its allocation in its mandate, with its remaining exposure behaving like US Mid Cap Growth and Non-US Growth. From a "style" standpoint, this fund was 100% allocated to the "growth" factor.

When we apply these fund exposures to the portfolio's allocation to each fund, we see the portfolio's allocation to the markets.

ACTIVE WEIGHTS DRIVE TRACKING ERROR

The portfolio's active segment weights are the difference between the effective exposures and stated exposures. For example, the US large cap segments are both underweight their target exposures by about 3-1/2 percent. These active weights drive the tracking error that is structural in nature. The residual tracking error is the result of investment selection.

The portfolio reflected an excess return of 187 bps with tracking error of 242 bps. This is typically assumed to be the result of investment selection, but we see that this is false. The benchmark misfit produced 5 bps of excess return while contributing 79 bps of tracking error - quite an inefficient process, as we would expect of any unmanaged factor. However, it's more unsettling to see that this unknown factor was responsible for one-third of all of the portfolio's active risk.

WHERE'S THE EFFICIENCY?

Efficiency results from equalizing the percentage contributions to both risk and return for each portfolio segment. We create an "efficiency statistic" that is the difference between these contributions.

We easily see that benchmark misfit contributes less than 3% of active return while contributing almost 33% of active risk, and producing a substantial, double-digit inefficiency of -30 percent. This inefficiency is offset by the good news of an efficient fund selection process which contributed 97% of the active return and only 67% of the active risk.

FINDING THE SOURCE(S) OF THE PROBLEM

It's not surprising to see that unmanaged benchmark misfit risk is a significant source of tracking error, but this conclusion is not specific enough to help us solve this problem. We need to drill down to the fund level to find the source of this misfit risk.

In the case of this portfolio, there is a single substantial detractor from efficiency: the high-quality bond segment, which was filled by an aggressive, alpha-seeking "core plus" strategy that followed a familiar (but misguided) strategy of "loading the boat" with spread product and drifting very far from the risk profile of its AAA-rated benchmark. This typically involves underweighting Treasuries and other AAA-rated securities while holding lower-quality corporate bonds (BBB and split-rated "crossover" bonds) as well as high-yield and even emerging markets debt.

By applying this benchmark misfit analysis, we instantly identified the bond fund as the single problem that needs attention.

We now see that the bond fund's mismatch with its benchmark exposures accounted for half the tracking error in the portfolio (122 bps out of a total tracking error of 242 bps.)

WORTH THE RISK?

This benchmark misfit risk approach can help prevent the two problems we often see in actively-managed portfolios:

• Firing your best active contributors
• Holding onto your worst active detractors

This bond fund "Bad Boy" outperformed its benchmark by 187 bps, which is an impressive level of excess return - so long as you: a) ignore risk and b) look at the fund in isolation (vs its role within the portfolio.) We found that 305 bps of this spurious "excess return" came from its mismatched strategy - essentially the result of comparing its performance to the wrong bond benchmark. After adjusting for its true market exposure, we saw that its true "selection effect" was -118 bps!

In the context of each fund's contribution to the portfolio's active efficiency, we immediately see that the bond fund is the downside outlier, contributing almost twice as much to active risk as it contributes to active return. The portfolio's other funds lie reasonably close to the "Efficiency Line" where (in a perfect world) all assets contribute equally to return and risk, lying perfectly on the line.

"So, NOW what should we do?"

Armed with these insights, a portfolio manager can quickly identify whether the portfolio's underlying funds form a good "fund team" that complement each other's periods of weak performance while also helping to offset each other's drift from their mandates. For example, a US equity manager buying non-US stocks would pair well with a non-US manager buying US stocks. Or the "top-level" portfolio manager may underweight the portfolio's exposure to emerging markets funds, knowing that the developed foreign equity fund also invests in EM stocks. We can also apply the idea of "alpha diversification" where funds with complementary alpha patterns are selected for the fund team. (We dealt with this portfolio construction technique as a separate topic, which we have published in the article "Fund Evaluation from a Portfolio Perspective," Journal of Performance Measurement, Spring 2022.)

The investment firm's CIO office and the client's investment committee and board of directors should require this level of risk analysis to help them carry out their fiduciary responsibilities. Quite frankly, one has to ask the question: "How could you NOT know whether this is going on in your portfolio?"

Those "first adopters" of this approach will define the new "Gold Standard" for true portfolio construction and more meaningful fiduciary oversight.

"What's in YOUR portfolio?