Stay Connected!
ETF.com News Daily
ETF.com News Weekly
Sign up to ETF.com's newsletters.
U.S. Edition
Search Ticker
Journal of Indexes

Dealing With The Active

Related ETFs: DON
Share:

By William O'Rielly, CFA and Michael Preisano, CFA

Granted that active managers cost more for many reasons than passive managers and on average have performed worse, what can you do about it? The authors describe a straightforward and reasonable selection and allocation process that does much to minimize the costs and deal with the apparent underperformance. Your own average active manager doesn't have to be merely an average active manager, and you don't have to use him at all.

As the popularity of passive investing grows with institutional and retail investors, consultants need to continually reexamine the arguments in favor of and against passive investing. When addressing the active versus passive debate, it is important to determine which index is being compared and why that index may differ from active management styles. Under or over performance may have as much to do with an active manger's decision to own or not to own the top 10 index names as it does his real abilities at security selection.

Regardless of the top-heavy and other seemingly non-intuitive ways indexes are constructed, passive investing via index funds has long been popular with institutional plans and retail investors. A1998 Pensions and Investments survey of the top 50 leading index fund managers found that index assets in total had grown to approximately $1.4 trillion, a doubling since 1995. At the retail level, the Vanguard Index 500 Fund has grown to become the second most popular fund in the US as of August 1999. The fund then had $91 billion in assets, closing in on Magellan and $33 billion larger than the next largest fund. Of the $2.3 trillion dollars invested in open-ended US Equity funds as of August 1999, approximately $210 billion was managed to an S&P 500 mandate. This represents a growing trend at the retail level.

 Institutional and retail investors give similar reasons for selecting index investments. The most common include:

• Index funds outperform the large majority of active managers

• The index represents an efficient portfolio

• Indexing is a fully invested strategy

• Indexing is a low-cost method of participating in the equity markets

The last two reasons are valid and need to be effectively addressed by consultants. In fact, if consultants address these last two objections with proper manager research methods, the first two objections become less burdensome. A large percentage of the underperformance and seeming inefficiency of large-cap US equity fund managers during the last 10 years can be attributed to fees and cash balances (see preceding article). Using Morningstar data, the table below shows an arithmetic return average of 271 non-index large-cap, US equity funds with 10-year track records. The 271-fund universe was a reasonably homogeneous universe of larger company US equity funds.

The relative under or over performance has been decomposed into an expense, cash and stock selection effect as follows:

As the table shows, in any given year, active management fees subtracted for an average of 1.14% from the managers' benchmark-relative performance. An average cash balance of 6.75% also 'dragged' performance down.1 The upward trending market during the ten-year study period resulted in an average 'cash drag' of 83 basis points in this sample. When management expenses and cash balances are stripped away, an approximation of gross-of-fee, equity-only returns vs. the benchmark shows just 52 basis points is left as a security selection effect.

TABLE 1: AVERAGE FUND vs. LARGE CAP BENCHMARK
  Average Over/ Avg. expense Return lost due Security Selection
  Under Perf. = Ratio + to cash drag + Effect +
1989 -4.95 -1.14 -1.24 -2.45
1990 -0.98 -1.14 0.8 0.06
1991 3.61 -1.14 -1.94 6.66
1992 0.06 -1.14 -0.28 0.79
1993 2.02 -1.14 -0.61 2.62
1994 -2.80 -1.14 0.40 -1.97
1995 -5.63 -1.14 -1.78 -3.91
1996 -3.28 -1.14 -0.98 -1.44
1997 -6.47 -1.14 -1.46 -4.82
1998 -6.44 -1.14 -1.16 -3.71
Average -2.49 -1.14 -0.83 -0.52

This analysis could be expanded to quantify any small-cap, international or fixed income effect that may have also hurt 1989-1998 performance relative to the S&P 500. Survivorship bias may have influenced the results. However, these results do show that a research process that screened for lower-cost fully invested managers 'started off' with less of a disadvantage than a process that was insensitive to expenses and cash balances.

This process was repeated for small-cap and non-US equity funds using the Russell 2000 Index and the MSCI EAFE Index. The small-cap study looked at 217 non-index small-cap funds based on Morningstar category classification with at least a five-year track record. The periods from 1989-1994 used available historical returns from those small-cap funds that met the initial five-year performance screen.

As the table shows, the average fund outperformed the Russell 2000 Index by 142 basis points despite an average expense ratio of 1.37% and a 0.74% "cash drag." The average small-cap fund held 7.1% in cash, slightly more than the average for large-cap funds. The cash penalty in small-cap and international was not as great as the large-cap penalty since small-cap and non-us equities did not outperform cash holdings as widely as did large-cap US. The largest difference between equities the large-cap and small-cap study is the average security selection effect. The average large-cap manager subtracted 52 basis points from the index return compared to the average small-cap manager's outperformance of 353 basis points.

TABLE 2: AVERAGE FUND vs. RUSSELL 2000
  Average Over/ Avg. expense Return lost due Security Selection
  Under Perf. = Ratio + to cash drag + Effect +
1989 5.39 -1.37 -0.94 7.70
1990 8.73 -1.37 1.31 8.79
1991 -0.58 -1.37 -2.84 3.63
1992 -3.58 -1.37 -0.80 -1.41
1993 -2.32 -1.37 -0.96 0.01
1994 1.29 -1.37 0.35 2.31
1995 2.66 -1.37 -1.81 5.84
1996 3.43 -1.37 -1.05 5.85
1997 -1.05 -1.37 -1.14 1.46
1998 0.24 -1.37 0.51 1.10
Average 1.42 -1.37 -0.74 3.53

The same study was also done for 171 non-index international equity funds with a five-year track record. Again if the fund had a longer than five-year track record, all available history was used to generate the annual year arithmetic average for 1989-1994. The results are as follows:

TABLE 3: INTERNATIONAL FUNDS' PERFORMANCE
  Average Over/ Avg. expense Return lost due Security Selection
  Under Perf. = Ratio + to cash drag + Effect +
1989 11.53 -1.36 -0.79 13.68
1990 13.12 -1.36 1.00 13.48
1991 0.67 -1.36 -0.44 2.47
1992 7.50 -1.36 0.44 8.42
1993 -4.40 -1.36 -1.94 -1.10
1994 -8.83 -1.36 0.29 -7.76
1995 -1.33 -1.36 -0.26 0.29
1996 6.74 -1.36 -0.44 8.54
1997 2.55 -1.36 0.03 3.88
1998 -7.48 -1.36 -0.44 -5.68
Average 2.01 -1.36 -0.26 3.62

The international results are similar to the small-cap study. During the 10-year period the average international mutual fund outperformed the index by 2.01% despite an expense ratio of 1.36% and a 26 basis-point cash drag. This study supports the theory that there are enough inefficiencies in international markets to support active management.

 Passive investing can also appeal to investors familiar with two interrelated academic concepts: the efficient market hypothesis and Capital Asset Pricing Model (CAPM). The efficient market hypothesis, in its strongest form, states that stock prices rapidly respond to all public and private information. No single group of investors has monopolistic access to these sources of information and therefore no single group can consistently earn above average profits. Security prices are an unbiased reflection of all currently available information, including a stock's risk. Therefore the expected returns implicit in the current price of a security should reflect its risk.2

In the 1950's and early 1960's Harry Markowitz derived a formula for computing the (risk) variance of a portfolio. This brought to light the importance of diversification as a way to reduce the overall variance of a portfolio. The efficient frontier was defined as that group of portfolios that had minimized variance for a given level of return. CAPM, using the concept of a risk-free asset, expanded Markowitz's work by defining a Market portfolio as the aggregation of all investments made by risk-averse profit-maximizing investors and by demonstrating that this Market portfolio was efficient. The "Market" portfolio theoretically includes all risky assets in proportion to their market value. In investment terms, a market-value (capitalization) weighted index serves as a proxy for the Market and represents the risk-averse investment community's aggregate valuation of risky securities.

These academic theories have been countered with a variety of empirical market studies, most notably the Fama and French study3, which demonstrate pockets of securities that do not seem to be efficiently priced. Size-effect studies, weekend-effect studies and neglected-firm effect studies, to name a few, all attempt to explain seemingly abnormal (non-efficient) return patterns of certain groups of stocks.

Are stock prices always efficient? Does the Market basket represent the best aggregation of these stocks?

Against the average manager they seem to. But no law says you have to hire the average active manager

Addressing these "academic" concerns, consultants must first identify active managers who can articulate a rational philosophy and demonstrate a consistent process that has enabled them to outperform a passive index or market portfolio. Many active managers themselves have done empirical studies to support their particular investment strategy. Secondly, consultants need to combine these active managers into a multiple active manager portfolio, such that each segment of the broader "Market" portfolio is represented: small- and large-cap, value, and growth. This multiple-manager approach leaves room for indexing particular segments of the overall portfolio if compelling active management alternatives cannot be identified. The multi-manager approach also allows enough flexibility to include specific style-consistent active management specialists who have demonstrated stock selection skills.

To empirically test the results of a price and style sensitive mutual fund screening approach, we set forth a list of preferred characteristics that candidate funds should meet:

MUTUAL FUND/PRODUCT SCREENS

• Is the performance history long enough to produce statistically valid analysis?

• At least a five-year track record.

• Is the current portfolio manager responsible for the performance track record?

• Manager tenure greater than five years.

• Is the fund's expense ratio reasonable, based on an average of similarly invested funds?

• Expense ratio is below average for investment style.

ASSET ALLOCATION CONSISTENCY

• Is the fund fully invested in the stated asset class?

• At least 85% invested in stated asset class.

• Is the fund diversified within the stated asset class?

• No greater than 40% in a given industry.

• Has asset turnover been reasonable?

• Annual turnover of assets has been no greater than 1.5 times the investment style average.

INVESTMENT STYLE CONSISTENCY

• Has there been consistency of the actual portfolio holdings of the fund with the stated management style?

• The fund's relative median market capitalization and median earnings multiple and book ratio are less than 25% different from the stated investment style.

• Has there been consistency of the fund's returns-based style analysis with the management style?

• The fund's analysis shows at least a 50% exposure to the stated investment management style over independent three-year time frames.

PERFORMANCE

• Is the fund competitive on a risk-adjusted basis to similarly invested funds?

• The fund's Sharpe Ratio is greater than that of the style index over three- and five-year periods.

• Is the fund adding excess return over and above a risk-adjusted style benchmark?

• When compared to a "blend" of investment style benchmarks over a moving time frame, the fund shows competitive return.

• Is performance being driven by one spectacular year ?

• Fund shows competitive calendar performance relative to a style index.

• How has the fund performed in down markets?

• Fund does not fully participate-tends not to decline as much as market-in market declines.

Beginning in 1997, we selected three to five Select funds in each of five equity styles, US Large Value, US Large Growth, US Small Value, US Small Growth and Non-US Developed Country Equity.

To measure how effective this fund selection process has been, we took the arithmetic average of the Select fund performance from each of the five styles and compared the annual performance for 1997 and 1998 to both the Morningstar Category Average and a representative benchmark.4 Most of the Select averages outperformed the category average from Morningstar.5

TABLE 4: SELECTED FUNDS' PERFORMANCE
Annual Category Select List Morningstar Category Respective Index
Returns Average Average  
International '97 9.47% 5.05% 2.06%
International '98 17.03% 12.26% 20.34%
Large Growth '97 29.78% 25.41% 30.50%
Large Growth '98 36.72% 35.45% 38.71%
Large Value '97 29.47% 26.78% 35.18%
Large Value '98 13.02% 12.57% 15.64%
Small Growth '97 11.23% 16.03% 12.96%
Small Growth '98 2.93% 5.26% 1.23%
Small Value '97 30.67% 29.64% 31.79%
Small Value '98 -8.14% -7.20% -6.46%

This screening process was also tested at the portfolio level. The allocation below represents a typical global all-equity portfolio.

The table below represents the blended portfolio above implemented in three different ways: using the Select List Average fund, using market indices and using the Morningstar category average.

  Select List Morningstar Index
1997 23.86% 21.23% 24.59%
1998 17.81% 16.66% 19.85%

In both years, a blended portfolio using the average Select fund outperformed a portfolio constructed using the average Morningstar category fund. However, the Select portfolio underperformed the respective blended benchmark. During the past two years it has been very difficult for an active manager to outperform the large-cap index due to a somewhat small number of companies that have been driving the market. Any manager underweighted in that relative handful of securities was likely to look bad by comparison.

In addition, the Select small-cap growth and value performance slightly lagged due to the poor performance of a single fund in each style.

As previously shown, the average large-cap US equity fund has underperformed the S&P500 on an annualized basis over the last 10 years. However, a research process that identified reasonably priced active managers who had remained relatively fully invested, would have made up most of that underperformance. An effectively constructed multiple-manager portfolio can participate in the various sub-styles of the equity market while adding incremental value as each specialist attempts to outperform their style-specific benchmark.

Based on the results of this study, we developed an implementation that allows for active management of small-cap and non-US equities and a combination of active and passive management for large-cap equities. Within Large-Cap Equity specifically, a combination of active and passive management has historically shown favorable risk-return characteristics. We compiled rolling five-year risk and return figures for three different implementations of large-cap equity:

For active management returns, the Morningstar Large Growth and Morningstar Large Value Category Median returns from 1976-September 1999 was used. The S&P500 Index Total Return was used for the large-cap passive proxy.

An all active blend: 50% Large Growth / 50% Large Value / 0% Index
A combination active passive: 25% Large Growth / 25% Large Value / 50% Index
An all passive blend: 0% Large Growth / 0% Large Value / 100% Index

The rolling 5-year volatility figures do not differ appreciably from an all-active blend to an all-passive blend.

The fact that the average active large-cap fund has been underperforming the S&P over the last five years makes the current time frame argue for an all-passive large-cap solution. The screening process detailed in this discussion however has added value vs. the category average across most styles even during the last few years. A precise active-passive blend is difficult to quantify. Empirical trends suggest that a blend of both will moderate volatility somewhat and smooth out index outperformance and underperformance cycles. If future market movements are spread among a broader base of stocks, the underperformance vs. indexes exhibited in the past several years by active managers as a group may well shrink In such an event a careful selection process like the one outlined above could lead to outperformance.

1 The 'cash drag' was calculated by multiplying the cash percentage by the difference between the S&P 500 Index and the 90-day T-bill. The cash position was calculated by taking the average cash balance across all 271 funds since 1991. Trading costs were not explicitly considered here. Trading costs were built into the selection effect.

2 Frank K. Reilly, Investment Analysis and Portfolio Management, 4th Edition, 1994 The Dryden Press: 268-270.

3 Eugene F. Fama and Kenneth R. French, "The Cross-Section of Expected Stock Returns," Journal of Finance 47, no. 2 (June 1992): 427-465.

4 Russell 1000 Growth, Russell 1000 Value, Russell 2000 Growth, Russell 2000 Value, MSCI EAFE.

5Averaged set of given Select funds for 1997 and 1998.

 

Discussion

Post a Comment
Comment:
Name:
E-mail:
Home page: (optional)
CAPTCHA Image Reload Different Image
Type in the
displayed
characters:
Email follow-up comments to my e-mail address SUBMIT

IndexUniverse is excited
to announce our new name: