Blog Post

Asset Owner Portfolio Attribution | A Hierarchal Approach that Bridges Responsibility & Results

Related Articles
You're reading:

Asset Owner Portfolio Attribution | A Hierarchal Approach that Bridges Responsibility & Results

Alex Serman • January 15, 2025


CREATING A BRIDGE BETWEEN RESPONSIBILITY AND RESULTS

Performance measurement focuses on the calculation of investment returns, while performance evaluation explains the sources of those returns. The first activity (calculation) lives in the realm of facts, with clear definitions of the factors at work, and a consensus on the procedures that turn credible amounts into equally-credible returns. Those facts show the relative change in value resulting from the investment process.

The second activity (evaluation) lives in the realm of information – essentially a set of conclusions customized to a specific context or situation. The goal is to identify the sources of the changes in value of the investment portfolio. While performance measurement answers the “what” question (“What was my return for the past quarter?”) performance evaluation answers the “why” question: “Why did I earn this return?”

There are reasonable guidelines for performance evaluation and presentation, and applying them appropriately depends on the judgement of the person who creates this critical information. There is no single answer that explains every situation, but there is a credible approach that recognizes the unique aspects of each client portfolio. The key is determining which approach is most representative of the active investment process at work.

In this article, we move from the traditional approach to performance attribution (which usually deals with single-product funds, managed by a single decision maker) to examining the results of an asset-owner portfolio. This is typically a highly-diversified portfolio invested across several asset classes, with decision-making responsibilities made by several participants.

Our case study is a portfolio that includes global equity, global bonds, and alternative investments. Each of these asset classes is divided into its primary segments, creating a second level in an emerging hierarchy. Then, each of these segments is divided into its own unique groupings, completing our three-level hierarchy. Our performance evaluation examines the returns and the sources of return at each level of the asset allocation hierarchy. We then relate each level’s results to the total portfolio, providing the information that the asset owner needs to understand the set of decisions and the results they produced.

ASSET-BASED VS DECISION-BASED VIEWS OF THE PORTFOLIO

Our asset hierarchy can be seen as a natural expression of the assets themselves. It can also be viewed in terms of the decision makers operating at each level of the hierarchy. This dual view allows us to evaluate both the decision makers and the asset results simultaneously. We “give credit where it is due” as we evaluate each decision appropriately.

Consistent with traditional approaches to performance attribution, we break our decision process into:

  • where they allocated capital
  • which specific investments were selected.

These are the familiar “allocation” and “selection” effects.

OUR CASE STUDY’S ASSET ALLOCATION

This case study uses a 70-30 strategy, with a top tier of these major asset classes:

  • 50% equity
  • 20% alternative assets
  • 30% bonds.

Our second tier of assets splits the equity by domicile (US vs Foreign) and splits the bonds into high-quality (HQ) and below investment grade (BIG.) The alternatives are split between illiquid and liquid assets.

The third tier of the asset hierarchy completes the diversification:

  • US equity by size (large, mid, small) and foreign equity by developed vs emerging economies (EM,)
  • High quality bonds by US vs Foreign, and BIG bonds by market segment (HY, leveraged loans and EM,)
  • Alternatives by segment: illiquid by private equity vs hedge funds, and real assets by real estate vs commodities.

The broader structure of the asset allocation is not obvious from the detailed view that is typically used in key documents, such as the Investment Policy Statement (“IPS”) which describes the relationship between the investment manager and the asset owner. Here is a typical format, using the strategy in our case study. We show the long-term allocation (i.e. benchmark weighting) and the active weighting for the performance measurement period.

SEEING THE DECISION MAKERS WITHIN THE ASSET HIERARCHY

By taking a “top-down” approach to the asset allocation, we begin to see the alignment of the assets with the decision makers. For this initial performance analysis, the portfolio is tactically overweight in its growth components (public equity and alternatives) with a corresponding underweight to bonds.

Who made this decision? Clearly, this falls in the realm of the “top of the house,” which is the board of directors, working with the OCIO (“outsourced chief investment officer) or the overall portfolio manager. These are the decision makers with responsibility for aligning the portfolio’s strategy with the organization’s financial goals. This includes developing the long-term strategy, and making any opportunistic shifts in market exposure that aim to take advantage of higher short-term returns, while avoiding exposures and risks that are not expected to be compensated.

Some organizations are large enough to manage the investment portfolio internally. Examples are state pension funds, the larger college endowments, and private foundations. In these cases, the investment staff are typically organized by asset class, with a direct relationship between the “organization chart” and the investment portfolio. Alternatively, the asset segments may be assigned to specialists, either internally or to external investment firms, at any level of the asset hierarchy. The focus here is on assigning responsibility for the long-term and tactical asset strategy.

The second level in the asset hierarchy reflects the primary grouping of investments within in each asset class, along with the organization’s short-term outlook on these markets. This provides additional insight into the tactical decisions. We look at each asset segment as its own portfolio to evaluate the tactical allocation decisions.

For example, the portfolio’s leadership is enthusiastic about the public equity markets, with a 14% relative overweight to that asset class. And within that asset class – at the second level of the asset hierarchy – we see a positive view of non-US equity. Bonds were underweighted, with a tilt toward higher quality. Illiquid assets (private equity and hedge funds) are not typically able to be tactically traded, and so any changes in short-term weighting may be the result of changes in valuation since inception. The differences in its weighting relative to the benchmark are still factors that drive the attribution results, even though they might not be result of short-term allocation decisions.

Our tertiary allocations may be described as “style” within each of the secondary grouping of assets.

EVALUATING PERFORMANCE AT EACH LEVEL OF THE HIERARCHY

LEVEL ONE ATTRIBUTION

The macro asset class performance is driven by market exposure and the relative returns of the assets within each group. The performance measurement period is Q3 of 2023.

The portfolio outperformed its benchmark by 220 bps with 34 bps attributed to its tactical weightings, the remainder of the performance unexplained. This residual will be explained by the decisions made across the next two levels of the hierarchy.

We see a critical concept emerging: the bulk of the excess return was not a true “issue selection effect.” That would only be true if this were a single asset class portfolio, with a single set of allocation decisions. Since there are two additional levels of allocation decisions to be made, we know that the true investment selections will be made at the last level of asset groupings. For now, we consider this unexplained attribution result as the sum of the remaining active decisions; it is essentially an “active residual.”

We must “unbundle” the effects of the remaining allocation decisions to identify the true “issue selection effect.”

Key concept #1: the “active residual” at each level is what is left unexplained after calculating its allocation effect.

Our summary of level one attribution:

“Allocation effects were modest, and in line with the tightly-controlled limits on “timing the market” via large swings in exposure between the major asset classes. Equity and bonds contributed fairly-equally to the positive allocation effect. Most of the portfolio’s excess return came from active decisions at the second and third levels of the hierarchy, within the equity and alternatives segments.”


LEVEL TWO ATTRIBUTION

We evaluate the relative performance of each asset class, as though it were its own portfolio.

This is the information that the overall portfolio manager would use in a performance review with the group responsible for each asset class. For example, this could be three departments in a pension plan or an OCIO firm, or three separate investment companies hired to manage these asset classes.

Our attribution analysis explains the sources of excess return for each asset segment. Once again, we see a modest allocation effect, with most of the active returns contributed by the active decisions made at the next level in the performance hierarchy.

Key concept #2: the allocation effect assumes that the portfolio holds the identical market exposures as the benchmark for every asset segment – any structural differences are included in the “residual” effect.


RELATING ATTRIBUTION RESULTS TO THE TOTAL PORTFOLIO

We present “attribution at a glance” with a view of each asset class presented individually, and in terms of its contribution to the total portfolio. We relate each sub-portfolio’s individual performance results to the total portfolio by applying its tactical group weighting to its individual attribution results.

For example, global equity is a 57% position in the total portfolio, and so we apply this weighting to its secondary-level attribution effects. In the global equity asset class, we explain the 83 bps of residual excess return as 45 bps from US equity and 37 bps from foreign equity. We drill down to identify a modest allocation effect of -3 bps with 86 bps of residual active excess return. This active residual will be explained by active decisions at the third level of the hierarchy.


LEVEL THREE ATTRIBUTION

Our last leg of the decision hierarchy examines the performance of the six sub-groups in each asset portfolio:

  1. US equity by size (large, mid, small)
  2. Foreign equity by market (developed by emerging economies)
  3. High Quality bonds (foreign vs US)
  4. Below investment grade bonds (high yield vs leveraged loans vs emerging markets)
  5. Illiquid alternatives (private equity and real estate)
  6. Liquid alternatives (real assets: real estate and commodities)

Starting with 151 bps of the active residual in level two global equity, we fully attribute this across the five “flavors” of equity, with 82 bps contributed by US equity and 69 bps contributed by foreign equity. These contributions are further broken down by style within each equity sub-group. We weighted these level three results by the level two tactical allocations for each segment, to link the attribution between these two levels.

As noted, the level three attribution results are next related to the total portfolio by weighting them by the product of the two prior-level tactical weightings. This completes our explanation of the 186 bps of level one residual excess return.

Level Three Attribution Effect * Level Two Tactical Weight * Level One Tactical Weight = Total Portfolio Effect

This third level in the hierarchy accomplishes three important tasks:

  • Measures the benefit of the final allocation decisions
  • Identifies the true issue selection effect, and
  • Shows the true tactical positioning of the specific investments in the portfolio.

The total portfolio summary of excess return can be viewed as a snapshot of the three levels of allocation decisions along with the set of investment selections. These decisions contributed somewhat equally to our total value earned over the benchmark. We can assign these results to each decision maker, which is essential to a rigorous oversight process.

The three levels of active market exposure produced 115 bps of excess return, most of this coming from the level three assets. That is reasonable, since this level has the greatest volatility within its market returns. This level is also the source of selection return, since this is where the specific investments reside.

We also confirm that the first two levels of decision making are focused exclusively on tactical calls on the markets, and our hierarchical methodology identifies and quantifies the results of those decisions. This allows us to focus on the source of true selection benefit, after attributing the prior allocation decisions to the appropriate decision makers.

WHY THE TRADITIONAL APPROACH TO MULTI-ASSET ATTRIBUTION FAIL

The traditional approach to asset owner performance attribution is fine for situations where a single group makes every decision for the portfolio. However, that assumption is only relevant for a single-asset portfolio. When decision responsibility is spread across different groups, a more robust and representative attribution method is required.

In our case study, the non-hierarchical approach got the tactical positionings of the investments wrong 50% of the time. This serious misunderstanding came from attributing all tactical benefits to the decision makers at the investment product level, absorbing credit that was due to the prior asset allocators and market strategy decision makers. Sadly, this is a common mistake in performance evaluation.

We can easily visualize these errors using the first two equity assets:

Large cap equity appears to be equally-weighted relative to the benchmark. This is the opposite of the true result. Why? Because the equity asset segment is overweighted by about 10% (given its relative weightings at levels one and two of the hierarchy.) To maintain its benchmark weighting, these level three exposures should be increased by 10 percent. We also see that mid cap appears to be overweighted, but it is truly equally-weighted, since its exposure in the portfolio reflects this 10% higher asset allocation weighting across equity. These weightings can also be observed in the level three analysis.

We see that the total allocation values are correct, but the detail does not describe what happened, because it fails to describe the decision process. “Just because it adds up doesn’t make it right.”


UNDERSTANDING THE VALUE OF ASSET OWNER PERFORMANCE ATTRIBUTION

Asset owners bring the dual complexity of active, multi-asset strategies that are managed by a group of decision makers. It is essential for performance analysts to measure, evaluate and provide insight into the process and the results of the decisions made by the participants in the active process. These evaluations must serve all the players. This requires an evaluation at each level of responsibility as well as an explanation of how these individual results affect the total portfolio. Only a hierarchical approach to performance attribution satisfies these requirements.

We acknowledged the role and the scope of the responsibilities for each level of the portfolio’s asset hierarchy. We provided insight into the allocation decisions at each level, and then added the effects of the investment selection process. This approach provides the information and insights needed by all the participants, helping them to demonstrate their success in achieving their individual goals, along with their contribution to the overall portfolio results.

.

.

.

Written in partnership with Stephen Campisi