Alpha Unleashed
By Howard Present and Ron Santangelo

• Fact: Less than 20 percent of all mutual funds beat their passive benchmark over long time periods.2
• Fact: The investment process of large institutional investors when they purchase active equity managers differs dramatically from that of retail investors.

The first implied question is, would large institutional investors, with their depth of resources and analytical capabilities, continue to invest in active managers if they obtain a 20 percent success rate? Ostensibly, the answer is "no."

Therefore, it begs the second question: whether investors with less than several billion dollars can obtain better results from active managers than a long-term buy-and-hold strategy achieves. Our research suggests strongly that the answer to this question is "yes."

The 19 Percent Solution

With the exception of the largest institutional investors, the standard model for purchasing active managers is a heavily front-loaded research-and-evaluation effort, resulting in the selection of a single manager or mutual fund for representation of each asset class. Underlying this approach is an explicit objective of holding the mutual fund for an extended period of time. This latter bias towards a buy-and-hold strategy for active managers is reinforced throughout our industry by both traditional behavior patterns as well as economic incentives for the investment managers and the advisors or consultants responsible for delivering the investment to the end client alike. Unfortunately, evidence is mounting that this model for accessing active managers is suboptimal.

There have been numerous studies addressing the ability for the average mutual fund to beat its benchmark over longer investment periods. One of the more recent was conducted by Thomas P. McGuigan and appeared in the February 2006, edition of the Journal of Financial Planning.2 His research evaluated large- and mid-cap domestic equity mutual funds over 10-year rolling periods from 1983―2003. He found that, on average, only 19 percent of the actively managed mutual funds were able to beat their corresponding passive benchmark. This would suggest that a classic buy-and-hold strategy for purchasing active managers would yield a "success" only 19 percent of the time.

Figure 1


Unfortunately, human behavior makes even the 19 percent level hard to achieve. One of the most comprehensive research efforts on this topic can be found in Financial Research Corporation's compelling 2001 report entitled Investors Behaving Badly.3 In this study, they evaluated the reported returns obtained by mutual funds, and that of mutual fund returns adjusted to reflect the impact of investor buying and selling patterns. The latter was accomplished by adjusting performance based on new flows into each Morningstar investment category, resulting in flow-weighted returns. The analysis was conducted within Morningstar investment categories, and evaluated results from 1990―1999.

The results were significant. As can be seen in Figure 2, on average, the flow-weighted returns were 500 basis points less per annum than the returns that the mutual funds actually earned. This is a 43 percent reduction in overall performance.

Based on the research just cited, and other studies with similar results, it is not hard to understand the siren song of advocates of indexing:

1) less than one in five active managers can beat their passive counterparts net of fees over longer time periods;

2) there are a limited number of broadly available mechanisms for identifying the 19 percent of managers that can succeed a priori; and

3) the purchase and sell patterns of investors reduce the expected returns by nearly one-half. Therefore, why not play with odds in your favor and simply purchase (and hold) low-cost index funds?

Before we jettison the efforts of the vast majority of the investment management industry, a useful question to answer is whether the flaw is with the active managers themselves, or with the methodology that most investors employ when selecting the managers. Our evidence shows that one of the main culprits is the industry's commitment to mutual fund buy-and-hold strategy.

There is little debate that the entire industry is oriented to a front-loaded effort to the purchase of mutual funds. Whether a retail investor investing based on the guidance of her broker, or a midsize endowment or Pension Plan investing on the basis of its consultant's advice, the time spent on researching funds to support a purchase decision is a significant multiple of the time spent evaluating a proactive sale of a fund in advance of a performance failure.

Unfortunately, the evidence regarding the persistency of relative performance of active managers runs counter to this approach. Santangelo Research and Investment Management conducted a study that grouped U.S. equity managers into quintiles based on rolling three-year performance within their Lipper investment objective group.

The top-two-performing quintiles were then evaluated to determine which managers were able to remain in the top two quintiles of relative performance over the subsequent 3-year period. Of the 12 size and style categories, on average, only 10.8 percent of top-performing managers maintained their status over the ensuing three-year period. The best level of persistency was 19.1 percent for small-cap blend funds, and the worst was 3.6 percent for small-cap growth.

This evidence strongly suggests that an investment strategy of active managers that employs a targeted holding period of even three years does not align with the realities of active managers' relative performance characteristics.

Alpha Cycles

Investors hire active managers based on their expected ability to create alpha, and to outperform peer groups and passive benchmarks. For those managers that can create alpha, it is not delivered consistently every day. It ebbs and flows based on the manager's alignment with different market cycles, with the cumulative impact of substyles and substrategies, and even good luck. Successful managers need to deal with the inevitable diseconomies of scale within the investment process, with the personal challenges that come with professional and financial success, and the internal dynamics of keeping the investment team intact. In sum, there is cyclicality to the delivery of alpha from successful managers, which we term the Alpha Cycle.

We learn two things from this persistency research. First, we know that nearly 90 percent of top-performing active managers decay from one three-year period to the next. We also know that about 19 percent of all managers will outperform over a rolling ten-year period. Therefore, some of the managers that fell out of the top ranks over three years will reemerge as they move through their alpha cycle. A perfect example of this is the track record of the Legg Mason Value Prime Fund, managed by the legendary Bill Miller. Mr. Miller's fund has one of the truly stellar track records of the past 15+ years, but a closer look at the rolling 36-month and 12-month relative performance of the fund within the Morningstar Large Cap Blend peer group reveals an Alpha Cycle at work (Figures 4 and 5).

Figure 2


Since the first yellow-highlighted text doesn't say "Large- Cap Blend" is it OK for the 2nd and 3rd instances of yellow highlighting to say "LC"?

Figure 3


Anyone fortunate enough, or skilled enough, to have purchased the Legg Mason Value Prime Fund in 1990 and then have the fortitude to hold the fund throughout the inevitable Alpha Cycle troughs over the ensuing 17 years has been richly rewarded. There are, however, two challenges imbedded in this example.

Figure 4


First, the brilliance of Mr. Miller was much harder to identify in 1990 than it is today in 2007. There were clearly investment professionals skilled enough to identify him early on, but the visual clarity of hindsight makes it much easier to identify investment excellence. Second, the traditional sell trigger for active managers is an extended period of relative underperformance, and even this great long-term track record generated many windows to have triggered such a sale.

Figure 5


Understanding the Alpha Cycle sheds light on the negative impact of traditional human behavior when purchasing and selling active managers. The purchase decision comes after a manager has reached a relative peak versus peers (purchasing at the "high" of the Alpha Cycle), and then selling after an extended performance decline (selling at a "low" of the Alpha Cycle). Effectively, investors are employing a "buy high, sell low" strategy for active managers, which can materially reduce the natural returns that superior active managers deliver.

Figure 6

Figure 7

Sub-Styles And Their Impact On Alpha Cycle

Equities have a size (large-, mid- or small-capitalization stocks) and style (growth or value) component. Their predominant size and style is how a fund is characterized by the marketplace. Typically, U.S. equity funds are placed in one of nine size and style boxes.

Although size and style indications help differentiate managers to some degree, it does not account for the different strategies employed by the managers. These different strategies result in subasset class size and style differences. For example, large-cap managers have different market-cap focuses, with some only looking at the largest 100 companies and others focused on 500 or more. The range of market capitalizations can run from $400 million to $10 billion. On average, large cap managers that are more highly exposed to lower cap stocks are more likely to outperform when small-cap stocks are dominating, and vice versa.

Differences also occur by how value and growth are defined. Traditional style indexes use one or several valuation ratios (i.e., stock price to book value) to differentiate value and growth into discrete categories. However, in reality, a spectrum of companies is created from value to growth where companies can be defined as one or the other but have dramatically different characteristics or begin to overlap.

Value managers may be deep value managers, intrinsic value managers, traditional value managers or relative value managers. Differences in size and style lead to managers with different substyles and hence different Alpha Cycles. Other factors, among many, that can impact a manager's results are its exposure to companies with different credit quality and/or industry sector biases.

These strategy differences can result in certain size and style managers outperforming their peer group during different market environments. For instance, a relative value manager tends to perform better than other large-cap value managers when growth is outperforming value, such as in the late-1990s.

In 2001 and 2002, value started to outperform growth managers but economic growth was still below 3 percent. In this environment, the better-performing LCV (large-cap value) managers tended to be traditional value managers. Deep value (DV) managers continued to struggle as the soft economic environment caused many investors to question the deep-value companies' ability to meet their cash flow needs. In 2003, when economic growth began to accelerate beyond 4 percent, deep- and intrinsic-value managers in general outperformed their value peers.

Figure 7 is a little exaggerated in the actual results but is a good indication as to what one may expect from these differences in value substyles over time. DV tends to perform well when the economy accelerates in the early part of the cycle, and they are a major beneficiary again when M&A transactions heat up. Whereas, relative-value managers tend to outperform when earnings growth stabilizes, and again when it is decelerating below 10 percent.

Alpha Cycle Investing

The advocates of passive investing correctly point out the flaws with the traditional means of accessing active management. They also point out that their model will outperform the average active manager 80 percent of the time. But their approach also dismisses the entire population of active managers. We believe that there is a better model―Alpha Cycle^TM Investing. Alpha Cycle Investing is a six-step implementation process for active management that dramatically improves the likelihood of success because it is fundamentally in alignment with the realities of how skilled active managers add value. Alpha Cycle Investing demonstrates that the flaw is typically not with the active managers, but instead with the traditional mechanisms of accessing and implementing investment strategies utilizing active managers.

Interestingly enough, this investment model is not new. In fact, it is common behavior for the largest institutional investors. And it can dramatically improve the likelihood of success for investors―even those that don't have tens of billions of dollars to invest. As we will show, Alpha Cycle Investing can improve the 19 percent odds of beating an index to 67 percent or more for rolling five-year periods.

The reason that Alpha Cycle Investing has not been more often applied to the investment portfolios of smaller investors has more to do with historical implementation costs, operational considerations and potentially misaligned incentives for the participants in the manufacturing and distribution sides of the business. Fortunately, operational and logistical barriers are being eliminated by improved technology and industry scale. Meanwhile, the advent of mutual fund wrap programs and other fee-based pricing structures are better aligning incentives with investor goals and eliminating many of the transactional costs that had prevented acceptance of this approach previously.

Portfolio Effect: Multimanager Solution

In the investment process articulated above, the first two steps are not differentiated from traditional approaches. The first big divergence is with step #3: using a multimanager solution within the asset class. There are several important advantages to this approach. We have already introduced the concept that active managers have cyclicality to their relative performance. In most other investments that demonstrate performance cyclicality, the default implementation model is a portfolio of such securities. For example, no one would ever consider investing in the U.S. equity market by purchasing a single stock.

A small portfolio of active managers within a given asset class provides critical types of diversification:

• Investment process and the variety of substyles and other investment biases imbedded in every active investment process, plus diversification at the firm level.
• Manager-specific risk and the unpredictable events that can derail any quality investment team, including outside personal issues affecting the portfolio managers, diseconomies of scale impacting funds with a sudden influx of new money, changing team dynamics, and retention of key professionals.

Forced Sell Discipline

Given the lack of persistency in relative performance, a proactive process for removing managers from the investment portfolio is required. As previously mentioned, 90 percent of active managers are subject to a significant decay in relative performance from one three-year period to the next. It therefore follows that the anticipated or default holding period of top-performing active managers needs to be less than 36 months. Our research suggests that the inflection point in the relative performance declines is between 12 and 18 months. Alpha Cycle Investing builds in a formal review period to assess each manager within the portfolio every 12 months. If constructed properly, the portfolio of active managers should see a 20 to 50 percent rotation every year.

Another way to understand this model is to put it into perspective of the Alpha Cycle.

If Figure 8 depicts the Alpha Cycle, the optimal approach would be to purchase the manager at an Alpha Cycle low (position 2), and then sell as they are heading down the decay curve (position 3). While academically appealing, this approach will never pass muster in the harsh light of day. Instead, we advocate a variation on the theme that can best be described as "buy high, sell high." In Figure 8, this would result in a purchase decision at position 1, and a sell decision at position 3. The managers are selected based on their emergence of a relative cycle high, and then held until they begin to slide down the decay curve. Because there is a portfolio of active managers, this proactive selling can be accomplished with only a partial turnover of the portfolio each year.

Fortunately, there is a live performance model to evaluate the implications of Alpha Cycle Investing. In October 2007, Active Index Solutions, LLC began publishing a new set of public indexes, with daily values calculated and distributed by the American Stock Exchange. These indexes are the industry's first indexes to be constructed exclusively of a select number of actively managed mutual funds, and follow the six-step process described above.

There are currently 18 Actif index^TM indexes available covering 15 different asset classes (nine traditional domestic-style/ capitalization categories plus six sectors). Some of the track records extend back 10 years, and most have at least a 5-year track record. An evaluation of the rolling three- and five-year relative performance of these indexes versus both their benchmarks and Morningstar peer groups is shown in Figure 9. All data in Figure 9 reflects net asset value returns of the indexes

The results can be evaluated through two lenses. The first would be from the perspective of an investor that has already decided to invest in active managers. This category represents 70 percent or more of the net new flows into mutual funds in the industry. For these investors, this investment paradigm can increase the likelihood of success, if success is defined by outperforming a traditional passive benchmark, by more than 300 percent. When compared to peer groups of other active managers, the Actif indexes were above median nearly 70 percent of all three-year rolling time periods. The indexes placed in the top quartile versus peer groups 18.1 percent of the time, and resided in the bottom quartile only 4.9 percent of the rolling periods. This ratio of top quartile to bottom quartile is 3.7:1, and strongly validates that not only will this approach create persistency in outperformance, it also dramatically reduces downside risk.

The second perspective would be that of the ongoing debate between active and passive investing. By achieving a blended success rate of 67 percent, with many of the indexes achieving success scores even higher, this strategy justifies revisiting some of the underlying premises supporting the arguments in favor of passive investing.

A 2006 study in the Journal of Indexes (March 2006) entitled "What if Active Works?"4 took a look at the active versus passive debate from the home field of the firm that has come to symbolize indexing― Vanguard. The study looked at Vanguard's family of actively managed funds and compared them to the corresponding index funds. The conclusion: 70 percent of the actively managed funds beat their benchmark over the ten-year period, with the average level of outperformance equal to 1.47 percent per annum. Intriguingly, the 70 percent number for the Vanguard funds looked remarkably similar to the 67 percent result that was obtained by the Actif indexes. A key point here is that the vast majority of the Vanguard active funds are multimanaged and subadvised.

Figure 8


The Vanguard funds also have low fees, and our research emphasizes the importance of this feature. The 18 indexes have an average expense ratio of 1.09 percent. If you calculate the results gross of fees, the indexes would rank in the top quartile 58 percent of the time, compared to 17 percent for NAV returns. A focus on low-cost pricing is therefore critical.

Figure 9

Conclusion

The investment industry has been asking for quite some time whether or not active managers can outperform passive investment. What is interesting is that the evidence put forward supporting passive investment is actually answering the wrong question. The typical evidence is how traditional indexes outperform the average active manager within a peer group. They do. Repeatedly. In some ways, this is a rather meaningless statistic.

The more interesting question would be whether or not skilled active managers can outperform their benchmarks. With the insights from Alpha Cycle Investing, we are refining the question slightly further and asking whether a small portfolio of skilled active managers can repeatedly beat their benchmarks. We believe the answer to this question is resoundingly "yes," especially if investors are willing to group. They do. Repeatedly. In some ways, this is a rather meaningless statistic.

The more interesting question would be whether or not skilled active managers can outperform their benchmarks. With the insights from Alpha Cycle Investing, we are refining the question slightly further and asking whether a small portfolio of skilled active managers can repeatedly beat their benchmarks. We believe the answer to this question is resoundingly "yes," especially if investors are willing to engage in a strict sell discipline and rotate their "portfolio" of active managers.

Copyright 2007 F-Squared Investments, LLC. all rights reserved. "Active Index Solutions," "ActiFIndex" and "Alpha Cycle" are service marks of F-Squared Investments, LLC.