CUSTOMIZING YOUR FUND TEAM

Focus on Strategy and Alpha Goal

TURNING YOUR PLAN INTO A CUSTOMIZED PORTFOLIO

At the start of our investment process, we defined our client’s monetary goals and tolerance for mission-related funding risk. We also developed a target return to balance the client’s current assets with their planned withdrawals from the portfolio. Using forward-looking projections of market returns, we then created an asset allocation as one component of delivering that return. The other part of the overall client return is an active return enhancement to the market return from the asset strategy.

The first part of the active delivery mechanism is a platform of investments. This allows us to apply our “team of funds” approach, producing a high degree of confidence around an excess return that provides a surplus of portfolio value in good times that helps to sustain our client during down markets. This fund team is customized to each client’s asset allocation and their excess return target.

Our two-pronged approach is an innovation in an industry that views portfolio construction rather simplistically: as a collection of top-rated funds that are added to a portfolio based solely on their individual merits. This ignores the interaction between these funds. Our previous paper showed that “alpha diversification” is a critical factor in creating a portfolio that minimizes the tracking error that erodes confidence in the delivery of excess return.

We will demonstrate the merits of customizing the “fund team” around both the asset strategy and the client’s desired excess return. Our rigorous process provides a strong basis for developing a forward-looking, risk-adjusted estimate of excess return (or “alpha”) that plays a key role in client success. This is especially important for those times when forecasts suggest that market return alone may be insufficient in meeting a client’s return target.

CUSTOMIZATION VS CONVENIENCE: THE ROAD TO SUCCESS

The typical approach to fund selection identifies a set of recommended funds and applies this lineup to all client portfolios, regardless of differences in risk tolerance or alpha goals across these client segments. They simply adjust the fixed weightings for each fund so that the aggregate segment exposures match each client’s asset allocation. This is the quintessential “one size fits all” fund lineup.

However, we had already demonstrated the significance of “alpha diversification” and the benefits of this approach:

• Higher portfolio alpha
• Lower portfolio tracking error
• Higher confidence in delivering alpha
• Lower portfolio total return volatility (with certain sets of funds used in the team approach)

Our process begins with the factors driving successful portfolio construction (i.e., fund alpha and alpha correlation) and then brings in the key decision of fund selection and weighting. Fund weightings are optimized by considering the constraints of the client’s asset allocation and overall alpha target. This creates a holistic and cohesive approach that produces the optimal solution for each client’s unique goals. 

FRAMEWORK FOR OUR CLIENT PORTFOLIO SOLUTIONS

Our study included four asset allocations across six alpha targets. This produces 24 possible fund teams that we used to answer these critical questions:

• What were the results in terms of total portfolio returns and volatility?
• What were the active results in terms of alpha and tracking error? (Alpha is volatility-adjusted excess return.)
• Where did we find portfolio efficiency?

Perhaps our most interesting question was this:
“How much different are the fund lineups as we move across asset allocations and alpha targets?”

Rather than keep our readers in suspense, we will answer this last question first. We found that the differences in the selection of funds across these client mandates were “modest but meaningful.”

In practical terms, there were certain funds that were used across many of the strategies, but their weightings would differ. These “core funds” provided stability across the fund lineups. We also used a variety of “customizing funds” that provided the efficiency we required across all client solutions. There were differences in the degree of alpha diversification employed. These solutions provided “the best of both worlds,” because we achieved true customization without creating a dizzying array of outcomes. 

CREATING OUR CUSTOMIZED FUND TEAMS

To provide a sense of continuity within a consistent set of market assumptions, we began with the “70-30” allocation used in our prior papers on portfolio construction. We then pro-rated the segment weightings to match the macro asset allocations. This maintained consistency in the “stylistic” aspects of the set of four asset allocations: 50-50, 60-40, 70-30 and 80-20.

Our alpha targets began at 50 bps. Alpha was calculated as the portfolio total return minus its volatility-adjusted market return. This conservative approach lowers the excess return when the portfolio takes on more volatility than the benchmark, and compensates the portfolio when it takes a lower level of volatility risk. This reflects our “risk aware” approach to investing and evaluating investment performance.

The range of alpha targets was from 50 bps to 175 bps, in increments of 25 basis points. 

INITIAL TOTAL RETURN RESULTS

We begin our analysis by examining the reward the market provided to the benchmark for moving from the most conservative strategy (50% equity) to the most aggressive strategy (80% equity.) Perhaps surprisingly, during this 12-year period, the reward for taking on over 400 bps of additional volatility was only about 50 basis points of additional return. (We consider this to be “rather skinny.”)

This is the type of environment where a reliable active strategy is so valuable. Even our “entry level” alpha target of 50 bps matched the reward for moving from the lowest to the highest level of market volatility risk. This alpha enables conservative clients to meet their target return without taking on unnecessary levels of market risk.

Equally notable is the observation that these additional levels of return (from 50 bps to 175 bps) were achieved without additional volatility risk. Surprisingly, the level of total volatility decreased slightly because of the active process. This demonstrates our earlier comment of sometimes achieving “the best of both worlds.” We see that volatility and return were primarily driven by market exposure, with active exposure contributing positively to both performance measures.

ISOLATING OUR ACTIVE RESULTS

Our 24 active teams were selected to minimize tracking error while delivering the target alpha.

Our chart reveals several interesting results:

• Tracking error was closely related to total return volatility
• Alpha initially increased more quickly than tracking error (i.e., active efficiency increased with alpha target)
• “Efficiency Inflection Point” at 100-125 bps of alpha, as tracking error began increasing significantly

This active efficiency is further demonstrated in our chart of Information Ratio. Our “risk aware” Information Ratio uses alpha instead of naïve excess return in its numerator.

As noted earlier, active efficiency increased as alpha targets increased, resulting in rising Information Ratios - until the “efficiency inflection point” was reached, and these ratios declined slightly. We found it interesting that active efficiency seemed to be greatest in the more conservative strategies. One takeaway may be that conservative strategies might be better suited to reaching for higher alpha, especially when market returns are expected to be modest. This provides an opportunity for these more risk-averse clients to meet their return targets and accomplish their monetary goals.

“SAY IT WITH CONFIDENCE”

Clients “pay their bills” with total returns, which are the sum of market returns and alpha. Therefore, we establish our minimum return expectations with a high degree of statistical confidence, using standard statistical methods. The modest benchmark returns at 95% confidence show the benefit of alpha in managing “return deficiency risk.”

The team approach and the power of alpha diversification combine in a portfolio construction process that provides true active efficiency: a target alpha with a minimum of tracking error. As a result, our portfolio teams deliver a meaningful amount of alpha, even under downside scenarios. By comparing the charts of high-confidence total returns and alpha, we see that market return and active return contributed equally to the overall results. For example, at a 100 bps alpha target, the 95% confidence average total return is about 100 bps, with the average alpha component delivering about 50 basis points.

This also suggests that a successful portfolio construction process adds efficiency to the investment portfolio.

EFFECTIVE SOLUTIONS: A DIVERSE SET OF FUND TEAMS

How do asset allocation and target alpha constraints guide the fund selection process? We begin with a look at the effect of applying a 100 bps alpha target to the creation of a fund team for each strategy: 50-50, 60-40, 70-30 and 80-20.

We see that the funds chosen, and the weightings assigned to them do change as our alpha targets change. Each solution minimizes tracking error at the target alpha within each portfolio. These are not the “one size fits all” strategies that are observed in traditional portfolios. Rather, they demonstrate varying degrees of alpha diversification by the number of funds used, along with the weightings applied to the funds within a segment.

We observed the greatest degree of fund diversification in the large, developed-economy equity segments. The high quality bond segment reflected the greatest use of “non-core” funds that were key to maximizing alpha diversification across all funds.

The funds are ranked by Information Ratio within their respective asset segments. We provided these information ratio rankings along with the number of funds in each segment, so that the relative position of each fund is clear. For example, the large growth fund (LG3) is #8 out of eight funds – an easy evaluation. The high quality bond fund (HQ5,) the mid value fund (MV3,) and the high yield fund (HY2) are all rated #4; however, the bond fund has a relative ranking of 4-out-of-nine funds, while the other funds are “dead last” at 4-out-of-four funds.

Next, we examined the various fund teams for a 70-30 asset allocation, across all alpha targets (from 50 bps to 175 bps, in 25 bps increments.) Once again, we see that the funds selected - and the weightings assigned to them – both change as we assemble each set of fund exposures that minimize tracking error risk at each alpha target. It appears that the alpha target factor produced a greater variety across the funds placed in each fund team. That is, alpha target was the greater driver of fund selection, as compared with a single alpha target across the various asset allocations.

ADDITIONAL OBSERVATIONS

We selected our funds from a platform of 57 candidates across 11 asset segments in our strategy. Only 38 of these platform funds were used in our teams. Their inclusion was based on a combination of their individual performance along with their contribution to alpha diversification and overall team efficiency.

Our fund teams include all fund rankings by information ratio, from the highest to the lowest. Surprisingly, one of our “core” funds came in at last place (#8-out-of-eight) in terms of its individual information ratio. This demonstrates the importance of understanding how each fund contributes to the team. Sadly, this characteristic is completely ignored in the traditional approach to portfolio construction, which only considers individual fund performance relative to a benchmark.

How often were these funds used across the strategies and alpha targets? Across our four strategies and six alpha targets, the maximum potential usage is twenty-four times. We show each fund’s usage, along with its ranking by information ratio.

Within our 38 funds, only 7 (about 18%) were used in every one of the 24 portfolios in our study. Of those seven ubiquitous funds, two of them were not in the #1 information ratio group: these were in 2^nd and 8^th place. Once again, we see the importance of evaluating each fund in terms of its contribution to the portfolio’s alpha and tracking error. Using the traditional approach to fund selection, it is likely that many of our funds would have been excluded from consideration. The result would have been a set of inferior portfolios with poorer client results.

Only eleven of the thirty-eight funds (less than 30%) used in the team portfolios were ranked #1 in terms of their information ratio. The average information ratio of the funds used in the teams was three. Here is the distribution of the funds used in these optimized team portfolios:

FINAL NOTES – AND LOOKING AHEAD

This portfolio construction process might be reminiscent of the phrase: “Old wine in new bottles.” After all, we used traditional asset allocation principles and methods that were introduced by Markowitz over seven decades ago. That is the “old” part.

The “new part” is the application of these ideas to the active process of fund selection, producing an innovation in portfolio construction: the creation of “fund teams” that outperform the results of its individual team members. (For example, the portfolios used in this paper delivered information ratios that were over 3x higher than their individual funds.)

We have observed that “mean-variance optimization” generally lowers market volatility in an asset allocation strategy by as much as twenty-five percent. By comparison, portfolios using the “team approach” eliminate as much as seventy-five percent of the tracking error of their individual funds – retaining as little as twenty-five percent of active volatility! This makes intuitive sense, since alpha is generally idiosyncratic, and therefore more diversifiable – as compared with market returns which are generally linked to the same set of economic factors.

There are several other aspects of portfolio efficiency that we can use to demonstrate the power of active diversification. These are demonstrated by the innovative performance evaluation metrics that we will feature in our next article in this series on true portfolio construction.

.

.

.

Written in Partnership with Stephen Campisi.