Proprietary quant models may not be so exclusive
* Managers often rely on third-parties
* Use of outside research may cost investors
By Jessica Toonkel
NEW YORK, June 28 (Reuters) - Usually proprietary means exclusive. But when it comes to quantitative equity models that some mutual fund managers charge a premium for, proprietary is a matter of degree.
Even fund managers disagree about the definition of proprietary. To some, it refers to models they built with third-party data. Others use it to describe their own personalized version of someone else's quantitative investing model.
"Quite often managers will pitch us their process and they will claim it is differentiated but that is not always entirely the case," said Stephen Miles, senior investment consultant and manager of research in the London office of Towers Watson .
That grey area means that investors in these so-called quant funds may have to do more digging into their managers' process to make sure they are getting what they are paying for.
Quant funds generally cost slightly more than non-quant funds, according to Lipper. Investors may ignore the premium because they think a proprietary quant model with a secret formula gives them an edge.
The ambiguous meaning of "proprietary" became evident most recently at BlackRock Inc.
Earlier this year the New York-based firm changed the prospectus for its $3.7 billion Large Cap Series to show that the models for the funds were "generated by third-party research firms." Previously, the prospectus had stated that a "proprietary multi-factor quantitative model" formed the investing strategy for the three funds.
The funds' manager, Bob Doll, had always used third-party models. But he had customized them, "applied relative weightings," and added his own fundamental analysis to construct a portfolio, he said in a statement.
Doll is retiring from BlackRock Friday.
Some managers of quantitative equity mutual funds use third-party models and tweak them to some degree while others bring third-party data into their own models, experts said.
For example, J.P. Morgan Asset Management, the asset management arm of JPMorgan Chase & Co, uses earnings estimates from third parties and MSCI's Barra software to help assess and manage risk, although the rest of its model is constructed using its own tools.
"I would describe that as proprietary, but one could say it isn't," said Dennis Ruhl, chief investment officer of the firm's behavioral finance group. "There does tend to be a grey area."
There are typically three steps to a traditional, quantitative equity fund manager's process: choosing or building the models, determining the weightings of the portfolios after running figures through the models, and constructing the final portfolio based on the models and other analysis, experts said.
That last step is usually proprietary, but the first two steps can be outsourced to third parties, said Vadim Zlotnikov, chief market strategist at Bernstein Research, a unit of AllianceBernstein LP.
"If a manager thinks they are adding value through the portfolio construction, that is what they focus on," he said.
QUESTION OF VALUE
Some managers believe that it is difficult to provide differentiated returns by relying mostly on third-party models.
For one, third-party models used by fund managers may only work for so long, said Sandip A. Bhagat, a principal and global head of equities at The Vanguard Group. The research firms that create them typically sell them to multiple fund managers.
"There is nothing unique about your insights and if enough people have access to that concept, money will chase the idea to the point that the idea just stops ... providing good returns," he said. Vanguard uses proprietary models, said Bhagat.
For the past five years, long-only quantitative equity funds are down 2.38 percent, compared to non-quant equity funds, which have fallen 1.07 percent, according to Lipper.
Investors need to be informed clearly, said Russ Kinnel, director of mutual fund research at Morningstar Inc. "Investors need to know how their managers are using third-party research and what else is going into that process."
A manager relying largely on third-party models and research should charge less than those who are primarily relying on their own models, Kinnel said.
The average expense ratio for quantitative equity funds is 1.12 percent, compared to 1.04 percent for non-quant funds, according to Lipper.
What's more, third-party research and models are typically paid for with "soft dollars" - or through commission revenue rather than through direct payments.
"Ultimately, these payments are shaving returns off the funds," said Harold Bradley, chief investment officer (CIO) for the Ewing Marion Kauffman Foundation, who oversees his own proprietary quant investments for the foundation. "It is not transparent at all."
- Tweet this
- Share this
- Digg this