Overdiversification taking a toll on performance

Too many managers in an asset class could hurt investment performance, according to a new article in the CFA Institute Financial Analysts Journal.

"My team meets with a lot of institutional investors across the globe, and we've been doing a number of research projects with clients doing a lot of risk analysis on their portfolios," said Ric Thomas, lead author as well as global head of strategy and research, investment solutions group, at State Street Global Advisors, Boston. "They've been overdiversified and it kept coming up over and over and over again."

The white paper, "What Free Lunch? The Costs of Overdiversification," examines how too many external managers can create enough of a loss of active risk that the overall active portfolio fails to outperform the benchmark.

The paper uses Pensions & Investments' data as of Sept. 30, 2015, utilizing the data of 88 of the top 200 U.S. defined benefit plans that provided complete manager data. According to the paper, plans on average utilize 75 external accounts.

The number of accounts vary by plan size: 16 plans with more than $50 billion in assets averaged 163 external accounts; 13 plans between $25 billion and $50 billion, 82 accounts; 39 plans between $10 billion and $25 billion, 56 accounts; 18 plans between $1 billion and $10 billion, 40 accounts; and two plans with less than $1 billion in assets, six accounts.

The paper was co-authored by Shawn McKay and Robert Shapiro, senior investment strategists in the investment solutions group at SSGA. State Street Global Advisors had $2.67 trillion in assets under management as of Sept. 30, $2.12 trillion of which was in passive strategies.

"We looked at the equity space just to keep it simple because there's a lot of asset classes that investors can allocate to," said Mr. Thomas. "We found that the average large pension fund — let's call it $1 billion and higher — allocates to between ​ 12 to 30 active equity managers. That's the average.

"So we said if that's the average, what should be the right number and what does it mean for active risk? If you're north of 10 or 15 active equity managers, it's really hard to get active risk in your portfolio and you need some active risk to outperform," Mr. Thomas said.

While the study looked specifically at active equity managers, Mr. Thomas said the principle of overdiversification can apply to all asset classes.

The authors also created two ratios to help investors determine the optimal number of active managers in an asset class.

The FAR ratio — fees for active risk — is equal to the management fee divided by active risk, and the FAS ratio — fees for active share — is the total management fee divided by active share.

Active risk is defined as the standard deviation of the excess return over the benchmark. Active share is the percentage of stock holdings in a portfolio that differ from the benchmark.

Utilizing the equation, for example, if a plan hires nine uncorrelated money managers each with an active risk of 6%, then the resulting active risk for that overall portfolio would fall to 2%. A lower active risk equals a smaller chance of outperforming the benchmark, the authors contend.

"If you're very overdiversified," Mr. Thomas said, "you can have an expensive index fund and you may be paying high fees for zero outperformance. What are you getting for what you paid? Your fees don't really change, but your active risk falls pretty dramatically."

The FAR and FAS ratios can help determine an optimal number of external managers, according to the authors. For example, assuming a plan finds 25 basis points per unit of active risk (a 1% active risk level) acceptable and each underlying money manager runs a 4% active risk level, the optimal number of managers might be four.

The article does note that numbers are "likely to change depending on the objectives and beliefs for each plan." Plan executives need to figure out an optimal amount of active risk and then choose the number based on the acceptable level of active risk, the paper said.

Mr. Thomas said rather than burden a portfolio with too many active managers, he suggests fewer active managers and providing those remaining with more capital, in order to gain more returns. While there still exists for some plans the traditional system of simply creating an asset allocation, selecting active managers and organizing a staff around selecting active managers, he said more plans are constructing their portfolios around risk.

"All investing to us is active management, because especially from an institutional investor's perspective, they have return and risk objectives and they constantly have to think about those risk-return objectives," Mr. Thomas said.

Plans should move "away from security selection and more toward asset allocation and factor allocation and looking at some of the macro risks," said Mr. Thomas. "Thinking about the big macro factor risks and spending more of their time around tracking those instead of picking individual stocks."

Asset owners echoed some of the concerns of overdiversification raised in the paper.

At its Nov. 20 meeting, the Nebraska Investment Council, which oversees $26.2 billion in assets including the $11.8 billion defined benefit plan and $2.1 billion cash balance plan, restructured its global equity portfolio in those two plans to 100% active from 75% active/25% passive, but will not hire additional managers.

Michael Walden-Newman, state investment officer, said in an email, "We had this very conversation in light of the recommendations that came out of our holistic review of our equity portfolio. We decided to spend our active manager dollars in the global equity space and go 100% passive in our international non-U.S. portfolio. This meant large increases to our active global managers, but we wanted to keep the number static for the reasons (the paper) mentions."

In meeting materials for the council meeting, Aon Hewitt Investment Consulting cited "reducing the number of active equity managers reduces the potential that managers will take offsetting active positions, thus creating a 'closet index' with active management fees."

Robert Maynard, chief investment officer of the $16.4 billion Idaho Public Employee Retirement System, Boise, said in an email: "I tend to believe a lot of active managers are not worth it, and lean toward the views of the paper — but I know there are studies that can show otherwise."

However, Mr. Maynard believes the overdiversification issue is dwarfed by other considerations.

"I am agnostic on these issues, primarily because of all the matters in constructing and running portfolios, I find this issue one of the least important (or, more to the point, one of the areas that has the least potential positive impact) — and that issues of manageability, explainability, robustness of the entire organization, risk control, etc. dominate this issue)," he said.

"The added complication, however, will be large the more and varied the active managers are, and the cost of that complication ... outweigh the actual 'alpha' that may be generated," Mr. Maynard said.​

The CFA Institute Financial Analysts Journal article is available on the CFA Institute's website.​