JEFFERSON CITY, Mo. - The Missouri State Employees' Retirement System is in the midst of a $769 million experiment with concentrated portfolios.
With four external active domestic equity managers running portfolios of only a few stocks each and one internal active tactical strategy, officials of the $5.34 billion fund want to see if they can outperform its domestic equity benchmark, the Russell 3000.
"We don't need filler stocks," said Rick Dahl, chief investment officer.
"In reality, managers have a favorite 25 to 30 stocks. The rest of the stocks are in their portfolios to keep them in line with their benchmarks; they are filler stocks. We don't need filler stocks, and we don't need to pay active management fees on filler stocks. That is the thinking behind the concentrated portfolios."
So far, Mr. Dahl said, "we've had favorable results" with the concentrated portfolios. "But I'm not ready to proclaim we've found the Holy Grail after three years of results."
The managers are: AmeriCap Advisers LLC, New York, with $149.4 million; Capital Guardian Trust Co., Los Angeles, $154.3 million; Oak Associates Ltd., Akron, Ohio, $139.9 million; and Legg Mason Capital Management Inc., Baltimore, $152.1 million. The program's internal portfolio, run by Mr. Dahl, has $173.4 million.
The concentrated program depends on a number of key elements, Mr. Dahl said, including:
* With few exceptions, MoSERS guarantees each manager a five-year contract. "The message we are giving to the managers is, `We don't care how you perform in the next quarter or year,"' Mr. Dahl said. "The kicker is, if they don't beat the index over five years, they don't get paid anything. There is no base fee."
* Managers work on a performance-fee basis. "It helps managers focus on MoSERS' benchmark, not on the manager's benchmark," Mr. Dahl said.
* The system's internal investment staff and external consultant, Summit Strategies Group, St. Louis, work on an incentive compensation basis. "So a portion of the staff's and consultant's compensation is predicated on making good manager decisions. That helps align the interests" of the fund and those working for it.
* Trustees delegated a lot of discretion to staff, enabling the fund to move swiftly in making investment and allocation changes. The staff has discretion to hire and fire money managers.
* The system doesn't limit its choices in considering managers to those with only long-term histories and large amounts of assets under management.
"Historically we hadn't had success in active (domestic equity) management, which is why we gradually moved to indexing," Mr. Dahl said. Since 1997, MoSERS has run an internal S&P 500 index fund, now at $925 million, that has generated an annualized 25-basis-point alpha.
"If we wanted to continue hiring active managers, the approach we had used wasn't working," Mr. Dahl said. "We had to try a different approach."
"One of the primary drivers of the concentrated portfolios" is the fund's existing diversification. Mr. Dahl said it already holds some 1,250 stocks through the internally managed index fund, externally managed enhanced index portfolios and an externally managed small-cap and midcap portfolio.
"In the concentrated portfolios, we focus on return rather than risk," he said. "We know these portfolios may be more risky, but we are comfortable with the volatility because when you put the concentrated portfolios with the rest of the equities, that volatility is mitigated."
Of the concentrated managers:
* AmeriCap, using both macroeconomic top-down and fundamental bottom-up research, has 40 growth-at-a-reasonable-price stocks;
* Capital Guardian, using a bottom-up fundamental approach, has 25 to 30 GARP stocks;
* Oak Associates has 20 to 25 growth stocks, mainly in technology, health care and financials;
* Legg Mason has 35 relative-value stocks; and
* the internal portfolio, run by Mr. Dahl, uses a tactical approach, moving opportunistically between asset classes. It currently is moving to high-yield bonds, unwinding a successful run with real estate investment trusts. "If we can't find anything from a valuation perspective that would outperform the S&P 500, we will take a neutral position and just hold the S&P 500," he said.
So far, the concentrated program has outperformed MoSERS' benchmark, the Russell 3000.
The composite return of the concentrated portfolios, including two managers that have been replaced, is an annualized 3.1% (from its inception April 30, 1999, through March 31). That is 5.7 percentage points higher than the Russell 3000 return of -2.6%.
By contrast, MoSERS' total domestic equity portfolio, including the concentrated program, returned 1.4% during the same period; and the entire MoSERS fund returned 2.2%.
Too soon to tell
While it's too early to judge the results of the concentrated portfolios, Mr. Dahl said, "As we get more convinced we have the right process in place, we will move more money out of indexing."
Mr. Dahl credits some of the outperformance to the ability of MoSERS' staff to reallocate without limitation among the concentrated managers.
"We have made a number of successful timing calls that contributed to the performance of the concentrated portfolios," he said.
For example, "we pulled $100 million away from Oak in March 2000 and redistributed it to the other concentrated portfolios," he said, a move that enhanced the performance of the program by moving to other concentrated managers whose performance would turn out to be better. Then last September, MoSERS reallocated $40 million back to Oak, which then outperformed by a wide margin, before taking it back a few months later.
"Oak has given us a wild ride," Mr. Dahl said.
The concentrated portfolios' composite return includes managers no longer in the program to avoid survivorship bias, Mr. Dahl said.
Zak Capital Inc., Minneapolis, started in April 1999 but was terminated after Suzanne Zak, chief investment officer, died last June. Zak Capital was replaced by Legg Mason.
Southeastern Asset Management Inc., Memphis, Tenn., which also started in April 1999, was dropped after three months because officials there declined to agree to a new performance-fee structure when MoSERS sought to refine the initial one.
Rewards for excess return
Excess performance essentially is split 70% to MoSERS and 30% to the managers.
So the concentrated managers can earn 30 basis points if they achieve an annualized 100 basis points of excess return over the benchmark at the end of five years. A manager can earn 60 basis points if it generates 200 basis points of excess return and so forth, up to earning a maximum 150 for 500 basis points or more in excess return.
"The expectation is the managers will earn a fee equivalent to performing 200 basis points over the benchmark," Mr. Dahl said. "If we decide to fire a manager (without cause), we will pay the manager as if it earned 200 basis points over the benchmark, even if the manager is, in reality, 1,000 basis points under the benchmark."
Managers are paid regularly throughout the five-year period as if they have returned 100 basis points of excess return. At the end of the five years, MoSERS and the managers will settle up. Either the fund will have to pay the managers more or the managers will have to pay back whatever fee they failed to earn. Neither side will pay interest.