FundQuest provides technology and research to wealth managers, and currently administers approximately $36 billion on its platform.  They are a subsidiary of BNP Paribas.  The current study is an update to a previous study released in 2006.

Methodology

The authors used Morningstar data for approximately 16,000 actively managed mutual funds, with each share class representing a separate “fund.”  The data covered the 15-year period from April 1992 to March 2007, and was restricted to funds with at least three years of history, representing approximately $7 trillion of assets.  Morningstar classifies funds into one of 58 categories; for US equities, these are the familiar nine style boxes based on market capitalization and style.

A best fit benchmark for each fund was determined by comparing historical returns to 56 possible indexes:

The following metrics were computed for each fund:

• Traditional alpha– The authors determined alpha based on comparison to a broad market index, such as the S&P 500.
• Real alpha – The authors also measured the alpha against the best fit benchmark, which they describe as the “additional return truly stemming from the unique ability and skill set of the investment manager.”
• Exotic beta - Exotic beta refers to the “return derived from exposure to other systematic risk factors (such as credit risk, liquidity risk, volatility risk).”  The authors calculate exotic beta as the traditional alpha minus the real alpha.  A high exotic beta can be desirable, since it indicates an exposure to a risk that is uncorrelated with established asset classes but has a positive expected return.

These three metrics were computed for each category for 3-year, 5-year, 10-year and 15-year time periods, based on an asset-weighted average of the funds in the category.

The authors then determined three factors for each category:

• Active/Passive Recommendation – If the real alpha was “positive” (at least 50 basis points annually over the time period in question) for at least three of the four time periods of the study, and the fourth time period was considered neutral, the category was deemed a candidate for an active strategy.  Conversely, if the real alpha was “negative” (less than 50 basis points annually over the time period in question) for at least three of the four time periods, and the fourth time period was considered neutral, the category was deemed a candidate for a passive strategy.  Everything else, including those categories for which insufficient data was available, was considered a candidate for either a passive or active strategy.

• Tactical Asset Recommendation –When investors are considering asset allocation and portfolio construction, categories with high levels of exotic beta are candidates for overweighting.  Similar to the criteria for the active/passive recommendation, if a category exhibits a “positive” exotic beta (greater than 50 basis points for at least three of the four time periods in the study, and the fourth time period was considered neutral), then the category is considered a candidate for overweighting.  The converse condition implies underweighting, and everything else is considered neutral with respect to weighting.

• Percentage range of actively managed mutual funds in category which outperformed category benchmark – For each time period, FundQuest determined the percentage of funds outperforming the category benchmark (i.e., the benchmark used by Morningstar for that category).  These values were then averaged for the four time intervals (based on available data), and are reported at the quartile level (0-24%, 25-49%, 50-74%, and 75-100%).

Results

The authors report the above three factors for each of the 58 categories.  Below is a table of those results for the nine US equity style boxes:

For example, small cap value is a candidate for active management (i.e., it had positive real alpha for at least three of the four time periods) and is also a candidate for overweighting (i.e., it had positive exotic beta for at least three of the four time periods).  Between 25-49% of small cap value funds (averaged across the four time intervals) outperformed the Morningstar small cap value benchmark.

The authors conclude that “the general theme was that more efficient categories were more favorable to passive investing while less efficient (meaning smaller or less heavily researched) categories showed benefits from active management. There were a number of exceptions to this theme though and the explanations for those variations are not within the scope of this paper.”  The above data directionally supports this conclusion, insofar as small cap blend and value are the only two candidates for active management.

On the subject of exotic beta, the authors conclude “more exotic (less correlated to traditional stock and bond investments) categories were more likely to outperform the broad markets.”

Creation of Investment Products

FundQuest is creating a family of mutual funds based, in part, on the results of this research.  Each fund in the family will cover an investment category, and the degree of active or passive management in the fund will be based on the results of the study.  For the active portion, FundQuest will choose a ‘best of breed’ active manager; for the passive portion a low cost ETF closely tracking the category’s index will be used.  Initially, one active manager per fund will be employed, on a subadvisor basis, but that may change as assets grow.  The range of active management in each fund is likely to vary from approximately 35% (for fixed income and large cap) to approximately 95% (for emerging markets).  Expense ratios are expected to be capped between 130 and 180 basis points, and will vary by category.  Fixed income will have the lowest expense ratio and internationals the highest.  Expenses for US equities will be capped between 140 and 160 basis points, depending on market capitalization (with large cap having a lower fee).

Our Analysis

The authors have done a thorough job analyzing a topic that has been studied at a broad level by many other researchers, and their research sheds light on a specific question that has received much less attention.  The results of their study agree with what would be expected intuitively.  Less efficient investment categories (e.g., small cap) benefit more from active management than more efficient categories (e.g., large cap).  Significantly, the authors indicate the degree to which they recommend an active strategy and an over- or underweighting for each category.

Whether these findings can translate to a real economic benefit to investors and advisors will depend, at least in part, on the following considerations:

• Composition Bias – Since the authors rely on Morningstar’s classification of funds into categories, their findings are susceptible to biases in those classifications.  As we have noted in other articles (e.g., see The Predictive Power of Morningstar's New Rating System), a number of studies have shown that mutual funds are often mis-categorized.

• Share Class Issues – The authors treat each share class as a separate fund.  Thus, within a category, one share class might exhibit positive real alpha, while another (higher expense ratio) share class might exhibit negative real alpha.  This behavior is not a consequence of manager skill; it derives solely from the expense ratio of the share class.  Given this, we asked the authors whether the study can provide information about the skillfulness of active managers.  They responded that “the study actually attempts to measure the net value added by active managers after expenses, not just manager skill.”  In addition, the bulk of assets tend to be in less expensive share classes. Since they are weighting the real alpha calculations by the asset size of the fund (i.e., share class), this effect is minimized.  However, the calculation of the percentage of actively managed funds that outperform their benchmark is a “headcount” calculation, and is more susceptible to this bias.

• Benchmark Assignment – The authors assign a benchmark to each fund based on the best fit correlation of returns against 56 possible benchmarks.  Our concern is whether this can result in inappropriate benchmark assignments.  For example, Morningstar recently reported the best fit index for a number of emerging bond funds is the Dow Jones moderate portfolio index.  Advisors will need to assess whether this is appropriate considering the political, currency, and liquidity risks in emerging bond funds.

• Time Periods – The authors calculate values using trailing 3-, 5-, 10- and 15-year results, all with the same end date  As a result, more recent history heavily influences results, since it is used in all four calculations.  We are now in a nearly five year bull market.   Funds and categories that generate alpha in bull markets will benefit, and will have good 3- and 5-year results, which in turn will influence their 10-year results (50% of which comes from the 5-year result).   The author’s goal with this approach is to incorporate both up and down market cycles; advisors will need to assess whether this methodology achieves that goal.

Dave Blanchett and Craig Israelsen have an article in the November 2007 issue of the Journal of Financial Planning, “Active versus Passive: The Debate Continues,” which highlights many of the issues researchers face in the studies such this one.  Our analysis draws on the concepts they present.

With respect to the proposed investment products, we believe the key issues will be the reliance on a single subadvisor for the active component, and whether sufficient alpha can be generated, net of the expense ratio.  These issues are interrelated.  For US equities, assume the expense ratio is 140 basis points (the low end of the projected cap) and 65% of the fund is actively managed (the mid-point of the projected allocation for all funds).  In this case, the active component must generate a positive alpha of 215 basis points to cover the expense ratio.  Advisors must analyze the likelihood of a single subadvisor delivering this level of alpha.  This consideration will be magnified for those categories with lower active components.  If a category has an active component of 35%, but still carries an expense ratio in excess of 100 basis points, the hurdle for the alpha that must be delivered by the subadvisor will be considerably greater.

Display article as PDF for printing.

Would you like to send this article to a friend?