Global Macro Attribution: A Framework for Assessing Plan Sponsor Performance

by Christopher Porter
J.P. Morgan Investment Analytics & Consulting
christopher.n.porter@jpmorgan.com

How can a plan sponsor appraise the quality of their investment decisions?

When managing a pension fund, it is necessary for the plan sponsor to monitor the investment returns of the fund managers they employ, but it must also have a tool for assessing their own investment decision-making abilities. This is where a Global Macro Attribution (GMA) analytical framework can be used to assess how the investment return of a plan has been enhanced by the plan sponsor’s investment decisions. GMA can isolate performance return attributable to decisions such as whether to underweight/overweight asset classes, managers, and/or styles, versus the plan’s strategic benchmark.

We believe our plan sponsor clients can use Global Macro Attribution to enhance their portfolio optimisation decisions. This analysis can be easily implemented, as the inputs required to produce GMA are generally available in fund accounting and performance measurement systems. In addition, when a client employs an investment consultant as plan sponsor, GMA may serve as a valuable tool for assessing the performance of this agent to the pension plan.

The purpose of this article is to show how GMA can assist with the management of capital allocation decisions by identifying good or bad investment decisions in the plan’s portfolio. The model depicted in this article assesses the following plan sponsor choices:

• Risk-Free Asset Decision
• Strategic Asset Allocation Decision
• Investment Manager Style Decision
• Active Management Return
• Tactical Asset Allocation Decision

An illustrative pension plan of ABC Company, based on “Evaluating Portfolio Performance” of the CFA Institute Level III Curriculum, will be used to explain each of these decisions.

ABC Company Pension Plan

Assume the plan sponsor of ABC Company’s pension plan possessed $100 million in capital to allocate at the start of the analysis period and was subject to the strategic benchmarks represented in Exhibit 1.

Exhibit 1: Strategic Benchmarks

This GMA analysis was conducted over a one-year time period where the total plan return was 3.74%. Beginning with this information, we can begin to assess each investment decision in turn, starting with the Risk Free Asset decision.

Risk-Free Asset Decision: The plan sponsor could have decided to implement a risk-free asset allocation by investing the entire portfolio in the risk-free asset, thus earning the risk-free return. This will be the first decision the GMA captures.  In this case, the risk-free rate of return was 0.25% or $250,000 in dollar terms on $100 million in capital.

Strategic Asset Allocation Decision: The strategic benchmark is the asset class allocation calculated to achieve the plan’s long-run risk and return objectives. Hence, the second decision the model captures is the Strategic Asset Allocation decision, which leads to the return earned by investing to mirror the strategic benchmark. The strategic benchmark earned 3.80% in the period, implying the additional return derived from investing in the strategic benchmark instead of the risk-free asset was 3.55% (3.80%-0.25%) or $3,550,000 in dollar terms.

Investment Manager Style Decision: The third decision captured by the model is the decision to allocate the plan’s capital to various managers. The plan sponsor may have decided to allocate capital to managers with alternative investment styles to the strategic benchmark. The capital allocation for ABC Company is shown in Exhibit 2.

Exhibit 2: Investment Manager Style Decision 

As illustrated, the plan sponsor has decided to allocate the equity portion evenly between a large cap and a small cap manager, and choose a fixed income allocation of 80% in U.S. bonds and 20% in European bonds. As such, the plan sponsor has decided to adopt a different investment style to the strategic benchmark; therefore, we need to determine the effect of this decision on the plan’s investment return.

The returns stated in the right column of Exhibit 2 are incremental, so the -0.80% return for the total plan is relative to the strategic benchmark of 3.80%.  Therefore, a portfolio that invests its capital in line with the weights and mandates specified in Exhibit 2 would earn 3.00% (3.80%-0.80%) from this investment manager style strategy. The plan sponsor Investment Manager Style decision reduced the plan return by 0.80% or $800,000 in dollar terms.

A point to note is that up until now we have only used index returns and not actual portfolio returns to assess the plan sponsor decisions. This captures the fact that a plan sponsor does not manage the capital at an individual portfolio level but is in control of the hiring (and firing) of managers, and of the assignment of portfolio mandates. The actual portfolio returns are considered when we move on to discuss the returns of active management.

Active Management Return: The Active Management return is the only part of the plan investment return that the plan sponsor does not have direct control over. This is the investment return added by respective managers as a result of their active management. Exhibit 3 shows the active performance of each investment manager.

Exhibit 3: Active Management Return

In this example, the incremental return from active management was 0.5% or $503,000 in dollar terms.

Exhibit 4: Tactical Asset Allocation Decision 

Based on the plan sponsor’s market expectations, it took the Tactical Asset Allocation decision to overweight the equity asset class by 5% to 65%, and underweight the fixed income asset class by 5% to 35%.  The additional plan return achieved by making this Tactical Asset Allocation decision was 0.24% or $240,250.

Now that we have considered all the decisions captured by the Global Macro Attribution model, we can summarise the results in Exhibit 5.

Exhibit 5: Global Macro Attribution Summary 

Exhibit 5 illustrates that the five investment decisions sum to equal the total plan return of 3.74%. By far the largest contributor to the total plan return was the Strategic Benchmark Decision, which added 3.55%. The plan return could further be decomposed in a format such as in
Exhibit 6.

Exhibit 6: Global Macro Attribution (in 000s) 

From this analysis, we can see which decisions enhanced the performance of the portfolio, which style choices paid off, which managers added value, and whether the Tactical Asset Allocation was a winning strategy.

Using Manager A as an example, this mandate earned $1.479m in total. If the portfolio had invested in the risk-free asset it would have earned $75,000. The Strategic Benchmark (MSCI World) earned an extra $1.425m. Opting for a large cap style reduced the return by $600,000, though this manager added $465,000 in active return. Finally, the decision to overweight equity and give this manager more capital than stipulated in the strategic benchmark added $113,750 to the investment return of this fund.

Considering the GMA output overall, the following has been determined

• The Risk Free Asset added $250,000.
• The Strategic Benchmark contributed $3,550,000, the majority of the return for the total plan.
• The Investment Style choice detracted from the total return on three managers, which illustrates these styles did not outperform the strategic benchmark. The decision to assign a small cap mandate was the only one that enhanced return.
• All managers added active return except Manager B who failed to outperform the MSCI Small Cap.
• The TAA Decision to overweight equity was a winner and added $240,250. 

Conclusion

In summary, this article has illustrated the value of Global Macro Attribution by showing how a plan’s total returns can be attributed to the major investment decisions of the plan sponsor. This information can then be used to make more informed investment decisions in the future. A GMA can be run over multiple time periods to ensure stability in the assessment of decision making abilities. In addition, the inputs required for such a GMA analysis are generally available from a high-quality performance measurement system, and the results can be dynamically adjusted to incorporate additional asset classes, portfolio distributions, and changes to strategic, style, or tactical decisions.

To view the next article, Key Rate Risk: Looking Beneath the Surface of Interest Rate, click here.