寻求投资组合经理的技能：主动份额和跟踪误差作为预测阿尔法的手段

Perspectives

February 2012 Legg Mason Capital Management Seeking portfolio manager skill

Past performance is no guarantee of future results. All investments involve risk, including possible loss of principal.

This material is not for public distribution outside the United States of America. Please refer to the disclosure information on the final page.

INVESTMENT PRODUCTSNOT FDIC INSURED • NO BANK GUARANTEE • MAY LOSE VALUE

Batterymarch I Brandywine Global I ClearBridge Advisors I Legg Mason Capital Management Legg Mason Global Asset Allocation I Legg Mason Global Equities Group I Permal Royce & Associates I Western Asset Management

February 24, 2012

Seeking Portfolio Manager Skill Active Share and Tracking Error as a Means to Anticipate Alpha1

1.5%

1.0%

Source: Antti Petajisto, “Active Share and Mutual Fund Performance,” Working Paper, December 15, 2010.

• There is a logical case for active management, but the key challenge is identifying above-average portfolio managers ahead of time.

• Most statistics in the investment industry and corporate America fail the dual test of reliability and validity.

• Active share and tracking error are both reliable statistics, and research shows that funds with high active share and moderate tracking error deliver excess returns on average.

• The long-term trend has been toward lower active share, which makes it difficult for mutual funds to generate sufficient gross returns to offset fees.

Equity markets are generally considered to be informationally efficient, which means that all relevant information is impounded in prices. Because it is difficult for an active manager to generate returns in excess of that of the market after an adjustment for risk, a common prescription is to turn to passive management in the form of index funds. This is a sensible approach for many investors. But the case for passive management has logical limits. For example, in economics there is an idea called the macro consistency test, which asks, “Would this approach work if everyone pursued it?” The answer for passive investing is no. Some percentage of investors must be active in order to ensure that information is translated into prices. Indeed, recent research shows that active management enhances, and passive management reduces, the informational efficiency of stock prices.2

In 1980, a pair of economists, Sanford Grossman and Joseph Stiglitz, wrote a seminal paper called, “On the Impossibility of Informationally Efficient Markets.” 3 Their basic argument is that there is a cost to making sure that prices properly reflect information. If there is no return for obtaining and trading on information, there is no economic incentive to do so. They propose that “those who do expend resources to obtain information do receive compensation” in the form of excess returns. In-depth studies show that active managers do indeed generate gross returns in excess of those of the market.4 However, those returns are less than the fees that managers charge. Careful studies of active management also reveal differential skill—that is, luck alone does not explain the results in investment management and a small percentage of managers deliver positive excess returns after all costs.5

Of course the challenge is to identify skillful managers ex ante. There are two major approaches to assessing manager skill.6 The first relies on an analysis of prior returns. The effectiveness of this approach relies on extracting useful information about skill by considering a sufficiently long time period and by controlling for various factors, including the types of risks the manager has assumed and the systematic versus idiosyncratic risk exposure. Returns-based assessments rarely make sufficient and appropriate adjustments to distill skill from the reported results. Further, simulations show that even skillful managers—those endowed with an attractive ex-ante Sharpe ratio—can deliver poor returns for years as a consequence of luck.7 In other words, even skillful managers won’t always beat their benchmarks and unskillful managers can do well for stretches of time as the result of randomness.

The second approach to testing manager skill is look at the portfolio holdings and characteristics of managers. Characteristics might include a portfolio manager’s age, education, and the size of his or her fund. The focus on portfolio construction and holdings allows for a more precise assessment of a manager’s process. The focus of this discussion will be on active share, a concept developed by a pair of finance professors named Martijn Cremers and Antti Petajisto, as a means to increase the probability of identifying a skillful manager in advance.8

What Is the Characteristic of a Valuable Statistic?

The worlds of finance and investing are awash in statistics that purport to reflect what’s going on.

Statistics that are useful have two features: reliability and validity.9 Reliability means that results are highly correlated10 from one period to the next. For example, a student who did poorly on a test last week does poorly this week, and the student who did well last week does well this week.

High reliability and a large contribution of skill generally go together. Finance researchers use the term “persistence,” which is the same as reliability.

The second feature is validity, which means the result is correlated with the desired outcome. For instance, say a baseball team’s offensive goal is to score as many runs as possible. An analysis would show that on-base percentage is better correlated with run production than batting average

is. So an enlightened manager would prefer on-base percentage to batting average as a statistic of offensive production, all else being equal.

The returns-based approach skips the two steps of reliability and validity and goes directly to the results. It doesn’t pause to ask: what leads to excess returns? It just measures the outcome. This approach works in fields where skill determines results and luck is no big deal. For example, if you have five runners of disparate ability run a 100-yard dash, the outcome of the race is a highly reliable predictor of the next race. You don’t need to know anything about the process because the result alone is proof of the difference in ability.

The difficulty with using a returns-based approach to assessing skill is that there is not a great deal of reliability, or persistence, in measures of excess returns. Researchers do find evidence for modest persistence but only when returns are carefully adjusted to account for style factors.11 But correlations over short periods, say year-to-year, for alpha based on the capital asset pricing model (CAPM) are quite low.

This problem of low reliability applies broadly to any highly competitive field that is probabilistic. Results, and especially short-term results, cannot distinguish between a good process and a poor process because of the role of luck. So going directly to the results gives little indication about the quality of the decision-making process and the skill of the participant.

In contrast, the approach that considers the holdings and characteristics of the manager allows us to look at both reliability and validity. Now the discussion shifts a bit. The questions become: which measures of an active manager’s portfolio reflect skill and therefore reliability? For example, a manager may be able to control the number of holdings, risk, turnover, and fees.

Next, of the measures that are reliable, which are highly correlated with the ultimate objective of delivering excess returns? Are there measures that are both reliable and valid?

Active Share + Tracking Error = Indicator of Skill

Let’s now take a closer look at active share. In plain language, active share is “the percentage of the fund’s portfolio that differs from the fund’s benchmark index.”12 Assuming no leverage or shorting, active share is 0 percent if the fund perfectly mimics the index and 100 percent if the fund is totally different than the index.

More technically, 1 N Active Share = ∑ ω fund ,i − ωindex,i 2 i =1 where:

ωfund,i = portfolio weight of asset i in the fund ωindex,i = portfolio weight of asset i in the index

Here’s a really simple example. Say the index has 10 stocks, weighted as follows:

Index Holdings Position Weight Stock 1 20.0 % Stock 2 15.0 Stock 3 12.0 Stock 4 11.0 Stock 5 10.0 Stock 6 9.0 Stock 7 8.0 Stock 8 7.0 Stock 9 5.0 Stock 10 3.0 Total 100.0

Now let’s say we have a fund of 10 stocks, weighted as follows:

Fund Holdings Position Weight Stock 1 10.0 % Stock 2 0.0

Active share is the sum of the absolute values of the difference between the weight in the index and the weight in the fund, divided by two:

In this basic example, you can see that active share is the result of not owning, or weighting differently, the stocks in the index (see stocks 1-10) and owning stocks that are not in the index (see stocks 11-15).

Generally, an active share of 60 percent or less is considered to be closet indexing and active shares of 90 percent or more indicate managers who are truly picking stocks. For the past 30 years, active share has been declining steadily for the mutual fund universe in the United States. For instance, the percentage of assets under management with active share less than 60 percent went from 1.5 percent in 1980 to over 40 percent today.

The selection of an index as a benchmark is obviously crucial. Research by Antti Petajisto shows that of the roughly 2,500 mutual funds he analyzed, 38.6 percent used the S&P 500 as their benchmark. Weighted by assets, 56 percent of the funds use the S&P 500 as a benchmark.

Other popular benchmarks include the Russell 2000 (8.8 percent by number, 6.2 percent by weight), the Russell 1000 Growth (8.4 percent by number, 5.6 percent by weight) and the Russell 1000 Value (8.2 percent by number, 8.4 percent by weight). Exhibit 1 shows the most common benchmark indexes.

Exhibit 1: Most Common Benchmark Indexes13 60%

50%

40% Percentage of Funds Percentage of Assets 30%

20%

10%

0% S&P 500 Russell Russell Russell Russell Russell Russell Russell Russell Other 2000 1000 1000 2000 2000 Midcap 3000 Midcap Growth Value Growth Value Growth Value

Source: Antti Petajisto, “Active Share and Mutual Fund Performance,” Working Paper, December 15, 2010.

There are two basic ways to raise active share.14 The first is through stock selection, as our simple example shows. That means either buying stocks that are not represented in the index, or owning stocks that are in the index but at a position weight that is higher or lower than what is in the index.

The second way to raise active share is through systematic factor risk15, which is effectively betting on factors by overweighting or underweighting16 industries. For example, a manager who is bullish on an economic recovery might overweight industries that are economically sensitive, or a manager who is bearish might overweight defensive industries. Tracking error, the standard deviation of the difference between the returns of the fund and of the index, does an effective job in capturing systematic factor risk. Tracking error puts more weight on correlated active bets than active share does. You can think of active share as a complement to tracking error and a measure that adds value in explaining fund results. A full picture of active management incorporates both active share and tracking error.

To illustrate the difference between these measures of active management, Petajisto offers the following illustration. Say a portfolio has 50 stocks. If all of the overweight positions are in technology stocks that move together, then small active positions will generate high tracking error. The portfolio has high systematic risk.

On the other hand, say the index represents 50 industries with 20 stocks in each industry and the fund selects one stock from each industry but weights that stock at the same level as the industry. In this case, active share will be high at about 95 percent but the tracking error will be relatively

low. As Petajisto notes, “active share is a reasonable proxy for stock selection, whereas tracking error is a proxy for systematic factor risk.”

There is a clear relationship between active share and tracking error. When active share is low, tracking error tends to be low and when active share is high, tracking error tends to be high. But the data show some amount of variation. For example, funds with tracking error of 4-6 percent can have active shares of 30 percent to 100 percent, while active shares in the 70-80 percent

range can be associated with tracking errors between 2 and 14 percent. This range of values for each measure of active management shows why it is important to distinguish between the two.

Exhibit 2 shows a matrix that classifies 401 mutual funds based on active share and tracking error using data from year-end 2007. The funds were sorted into quintiles based on each measure, with one as the lowest value and five as the highest value. The cells represent the number of funds that fall into the intersection of each pairing of quintiles. For example, if you examine the top right corner, you can see that 47 funds, or 11.6 percent of the sample, are in the highest quintile of both active share and tracking error. These are funds that are very different than their benchmarks and that have results that vary quite a bit from their benchmark. In contrast, the bottom left corner shows that nearly 15 percent of the sample—59 funds—are in the lowest quintile of both active share and tracking error. These are index hugging funds. To be in the top quintile, a fund must have an active share higher than 91 percent and a tracking error higher than 5.3 percent.

Exhibit 2: 401 Mutual Funds Ranked in Quintiles Based on Active Share and Tracking Error

Source: Morningstar, FactSet, and LMCM analysis.

A common measure of portfolio performance is the information ratio, which is excess return divided by tracking error. There is a fairly linear relationship between large factor bets (i.e., the overweighting or underweighting of industries relative to the benchmark) and tracking error. So in order to have an attractive information ratio, those factor bets have to really pay off to compensate for the high tracking error. Broadly speaking, they don’t. So active managers are better off maintaining high active share through stock picking than through sector bets.

The appendix provides a more sophisticated numerical example of active share and systematic factor bets based on a fictitious index and fund.

Active Share and Fees

The downward drift in active share raises an important issue in the active versus passive debate. If you’re going to go passive, go passive. Buy a fund that charges low fees and closely mirrors the index. If you’re going to go active, find a manager with a good process and high active share.

Academic research shows, for example, that the best ideas of money managers generate excess returns.17

Active managers who have crept toward closet indexing have created a losing proposition— market-like returns accompanied by fees higher than index funds. Petajisto provides an interesting case with Fidelity’s flagship mutual fund, Magellan. Peter Lynch made the fund famous by delivering outstanding returns in the 1980s when he managed it, in part by sustaining high active share. When Jeffrey Vinik ran the fund in the early 1990s, it had an active share above 70 percent. However, the fund’s next manager, Robert Stansky, who took over in 1996, took the active share below 40 percent and kept it there for a half dozen years.

Here’s why it’s so hard to beat the market with low active share and an average expense ratio.

Say a fund has an expense ratio of 125 basis points—roughly the average of all funds in Petajisto’s study—and the active share is 33 percent. That means that two-thirds of the portfolio

is earning the same return as the benchmark index. Therefore, the active part has to make up for the difference with massive outperformance. For example, in order to equal the benchmark’s returns, the active portion needs to earn an excess return of 375 basis points:

Gross return 1.25% Less expenses -1.25% Net return = 0%

Petajisto did find that low active share funds tended to have lower expense ratios, but the fees were sufficient to make outperformance extremely difficult. Generating excess returns is challenging enough, but low active share and average fees make the task even more daunting.

Reliability and Validity

We noted before that the returns-based approach to assessing skill skips the steps of reliability and validity. Now we apply the features of a useful statistic to active share and tracking error. The way to test reliability in this case is to examine the coefficient of correlation, r, between the active share for the same fund over two different time periods. Using a sample of approximately 400 mutual funds, the r between active share for 2007 and 2010 is 86 percent. (See exhibit 3, left side.) Cremers and Petajisto also found active share to be reliable. This makes sense, because active share is within the control of a portfolio manager.

Tracking error, too, appears to have good reliability. For the same funds and same time period, the coefficient of correlation, r, is 76 percent. (See exhibit 3, right side.) To put these figures in context, three-year alphas based on the CAPM have a coefficient of correlation of close to zero.

Exhibit 3: Reliability of Active Share and Tracking Error

r = 0.86 r = 0.76 1.00 25.00

2010 Active Share 2010 Tracking Error

2007 Active Share 2007 Tracking Error

Source: Morningstar, FactSet, and LMCM analysis.

The second feature of a useful statistic is validity—it leads you to the outcomes that you seek.

The research suggests that high active share is desirable if it is the result of stock picking, but less desirable if it is the result of factor bets that show up as high tracking error. Petajisto sorts 1,124 funds into five categories based on results from 1990 through 2009 (see exhibit 4). Stock pickers are in the highest quintile of active share, but in the bottom 80 percent of tracking error

within that active share quintile. He labels “concentrated” the funds in the highest quintile of active share that have the highest tracking error. These funds have high systematic factor risk. Funds in the stock pickers category have an average active share of 97 percent with an average tracking error of 8.5 percent. The concentrated funds have similar active share, at 98 percent, but have a tracking error of 15.8 percent, which is almost double that of the stock pickers.

Exhibit 4: Fund Categories and Their Statistics (5 = Top, 1 = Bottom)

Source: Antti Petajisto, “Active Share and Mutual Fund Performance,” Working Paper, December 15, 2010.

Funds that make relatively large factor bets are in the bottom four quintiles for active share but have among the highest tracking error within those quintiles. Moderately active funds are in the middle quintiles for active share but have tracking error that is much lower than the funds making factor bets. Finally, the closet indexers have low active share and low tracking error.

The test of validity is whether these categories correlate with excess returns. Exhibit 5 shows the results from Petajisto’s research. Importantly, these results include a period when large capitalization stocks did well (the 1990s), small capitalization stocks outperformed (the 2000s), and the financial crisis. Petajisto shows that stock pickers generate annual alpha of 1.39 percent, whereas all of the other categories have negative alpha. The funds that rely on factor bets are the worst performing category. However, funds with high tracking error and high active share perform less poorly.

Exhibit 5: Four-Factor Alpha for 1,124 Funds from 1990-2009 Based on Fund Category 1.5%

1.0%

Source: Antti Petajisto, “Active Share and Mutual Fund Performance,” Working Paper, December 15, 2010.

Petajisto’s large sample over 20 years provides a solid basis to establish validity. The smaller sample we used to establish reliability provided similar results. The funds that were in the highest quintile for active share and the bottom 80 percent of tracking error—64 funds altogether— generated annualized alpha of 3.8 percentage points from 2008-2010, well in excess of the results of the sample of all 400 funds.

The conclusion is that a thoughtful combination of active share and tracking error reflects the essential features of a good statistic: they are reliable and valid. Still, active share needs additional testing before it can be declared sufficient to indicate ex ante skill. One specific concern is benchmark selection. Small capitalization funds, for instance, tend to have higher active share than large capitalization funds. This is because small capitalization indexes have more stocks, with a lower average weight, than large capitalization indexes do. So the favorable results from the analysis of active share may stem in part from the fact that small capitalization funds beat their indexes more often than large capitalization funds do.18

Summary

Investment management is a very competitive business, in part because there are so many bright and motivated people seeking to beat their benchmarks. As a result, randomness plays a large role in determining results in the short term and it is difficult to deliver excess returns over time. Still, academic research shows that some funds do better than chance would suggest, and that active managers beat their benchmarks before fees. The fundamental question is whether funds with a good chance of outperforming their benchmark can be identified in advance.

There are two approaches to assessing managers. The first examines past results. This approach can reveal information if the track record is sufficiently long and enough controls are put into place so as to ensure that the results are not the consequence of risk. But relying on results in any domain that is based on probability is inherently troublesome, because the challenge of sorting skill and randomness is daunting.

The second approach studies the characteristics and behavior of the manager in order to assess whether he or she has a good process. To be useful, any measure of performance must have two features: reliability and validity. Most manager assessments use statistics that are neither reliable nor valid. We argue that a combination of active share and tracking error provides a glimpse into process. Taken together, these statistics are also reliable and valid. Based on this discussion, we can arrive at four conclusions:

• There is a role for active management. Having some active managers is a logical necessity. Research shows that active management increases the informational efficiency of stock prices and that passive management makes prices less efficient. The key challenge is identifying above-average managers ahead of time.

• Reliability and validity. Most statistics in the investment industry fail the dual test of reliability and validity. This is true in corporate America as well. As a consequence, it’s important to break down investment results further in order to get a better handle on potential skill.

• The combination of active share and tracking error provides insight. Active share and tracking error are both reliable statistics. For example, for 400 mutual funds (2007 to 2010) the coefficient of correlation for active share was 86 percent while tracking error was 76 percent. In seeking validity, the goal is to identify funds with high active share (top quintile) that are not in the highest quintile for tracking error. High tracking error indicates sizeable factor bets, which tend to deliver poor returns. Research by Cremers and Petajisto, confirmed by our own smaller sample, shows that funds with high active share and moderate tracking error deliver excess returns.

• There is a long-term trend toward lower active share. The percentage of assets under management with active share below 60 percent—considered to be closet indexing—has risen from 1.5 percent 30 years ago to more than 40 percent today. Passive management makes sense for a great deal of investors. But the essential message is this: If you’re going to be active, go active. Don’t own a fund with low active share, because the chances are good that the fund’s gross returns will be insufficient to leave you with attractive returns after fees.

I’d like to acknowledge the intellectual contribution of Arturo Rodriguez, CFA. As always, Dan Callahan, CFA, added great value in all aspects of preparing this piece.

Endnotes:

1 Alpha is a measure of performance versus a benchmark on a risk-adjusted basis. An alpha of +1.0 means the portfolio has outperformed its benchmark index by 1%. An alpha of -1.0 means it has underperformed its benchmark by 1%.

2 Russ Wermers and Tong Yao, “Active vs. Passive Investing and the Efficiency of Individual Stock Prices,” Working Paper, February 2010; Rodney N. Sullivan and James X. Xiong, CFA, “How Index Trading Increases Market Vulnerability,” Financial Analysts Journal, forthcoming (see http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1908227; and Jeffrey Wurgler, “On the Economic Consequences of Index-Linked Investing,” Challenges to Business in the Twenty-First Century: The Way Forward, W.T. Allen, R. Khurana, J. Lorsch, and G. Rosenfeld, eds., forthcoming (see http://archive.nyu.edu/bitstream/2451/31353/2/4_essay_Wurgler.pdf.) 3 Sanford J. Grossman and Joseph E. Stiglitz, “On the Impossibility of Informationally Efficient Markets,” American Economic Review, Vol. 70, No. 3, June 1980, 393-408.

4 Russ Wermers, “Mutual Fund Performance: An Empirical Decomposition into Stock-Picking Talent, Style, Transaction Costs, and Expenses,” Journal of Finance, Vol. 55, No. 4, August 2000, 1655-1695.

5 Robert Kosowski, Allan G. Timmerman, Russ Wermers, and Hal White, “Can Mutual Fund ‘Stars’ Really Pick Stocks?” Journal of Finance, Vol. 61, No. 6, December 2006, 2551-2595; Laurent Barras, Olivier Scaillet, and Russ Wermers, “False Discoveries in Mutual Fund Performance: Measuring Luck in Estimated Alphas,” Journal of Finance, Vol. 65, No. 1, February 2010, 179-216.

6 Russ Wermers, “Performance Measurement of Mutual Funds, Hedge Funds, and Institutional Accounts,” Annual Review of Financial Economics, Vol. 3, 2011, 537-574.

7 David L. Donoho, Robert A. Crenian, and Matthew H. Scanlan, “Is Patience a Virtue? The Unsentimental Case for the Long View in Evaluating Returns,” The Journal of Portfolio Management, Fall 2010, 105-120.

8 K. J. Martijn Cremers and Antti Petajisto, “How Active is Your Fund Manager? A New Measure That Predicts Performance,” Review of Financial Studies, Vol. 22, No. 9, September 2009, 3329- 3365.

9 William M.K. Trochim and James P. Donnelly, The Research Methods Knowledge Base, Third Edition (Mason, OH: Atomic Dog, 2008), 80-95.

10 Correlation refers to the relationship between two sets of data. When asset prices move together, they are described as positively correlated; when they move opposite to each other, the correlation is described as negative. If price movements have no relationship to each other, they are described as uncorrelated.

11 The most widely used are Carhart’s four-factor model (CAPM, size, valuation, momentum) and the Fama-French three-factor model (CAPM, size, valuation). See Mark M. Carhart, “On Persistence in Mutual Fund Performance,” Journal of Finance, Vol. 52, No. 1, March 1997, 57-82 and Eugene F. Fama and Kenneth R. French, “Common Risk Factors in the Returns of Stocks and Bonds,” Journal of Financial Economics, Vol. 33, No. 1, February 1993, 3-56.

12 Antti Petajisto, “Active Share and Mutual Fund Performance,” Working Paper, December 15, 2010.

13 S&P 500 Index is an unmanaged index of common stock performance. Russell 2000 Index is an unmanaged list of common stocks that is frequently used as a general performance measure of U.S. stocks of small and/or midsize companies. Russell 1000 Growth Index is an unmanaged index of those companies in the large-cap Russell 1000 Index chosen for their growth orientation. Russell 1000 Value Index is an unmanaged index of those companies in the large-cap Russell 1000 Index chosen for their value orientation. Russell 2000 Growth Index is an unmanaged index of those companies in the small-cap Russell 2000 Index chosen for their growth orientation.

Russell 2000 Value Index is an unmanaged index of those companies in the small-cap Russell 2000 Index chosen for their value orientation. Russell Midcap Growth Index is an unmanaged index of those companies in the Russell Midcap Index chosen for their growth orientation. Russell 3000 Index is an unmanaged index of the 3,000 largest U.S. companies. Russell Midcap Value Index is an unmanaged index of those companies in the Russell Midcap Index chosen for their value orientation.

14 Eugene F. Fama, “Components of Investment Performance,” Journal of Finance, Vol. 27, No. 3, June 1972, 551-567.

15 Systematic risk is the risk inherent to the entire market or the entire market segment. 16 Overweighting refers to having a greater allocation to a holding, sector, etc than the index; underweighting refers to having a smaller allocation to a holding, sector, etc than the index. 17 Randy Cohen, Christopher Polk, and Bernhard Silli, “Best Ideas,” Working Paper, March 2009.

Also, Klaas P. Baks, Jeffrey A. Busse, and T. Clifton Green, “Fund Managers Who Take Big Bets: Skilled or Overconfident,” Working Paper, March 2006.

18 Robert C. Jones, CFA, and Russ Wermers, “Active Management in a Mostly Efficient Market,”

Financial Analysts Journal, Vol. 67, No. 6, November/December 2011, 29-45.

Appendix: Decomposing Active Share

On the left is a fictitious index. The list shows the 50 stocks in the index, which industry each company is in, and the weight within the index. On the right is our fictitious fund, which has 25 stocks. Note that the stocks in the fund do not overlap completely with the stocks in the index. In other words, the fund holds some stocks that are not in the index.

Energy Company 4 1.4% Industrials Company 5 1.3% Consumer Discretionary Company 3 1.3% Industrials Company 6 1.3% Consumer Discretionary Company 5 1.2% Consumer Discretionary Company 2 1.2% Consumer Discretionary Company 1 1.2% Health Care Company 2 1.1% Industrials Company 2 1.1% Industrials Company 4 1.1% Financials Company 1 1.1% Materials Company 1 0.8% Utilities Company 6 0.8% Materials Company 4 0.7% Materials Company 3 0.7% Materials Company 2 0.6% Materials Company 6 0.6% Utilities Company 2 0.6% Utilities Company 3 0.6% Utilities Company 4 0.5% Utilities Company 5 0.5% Telecommunication Services Company 1 0.4% Telecommunication Services Company 3 0.1% Telecommunication Services Company 6 0.1% Total 100.0%

The table below shows the calculation of active share by placing the index and the fund next to one another. This fund has an active share of 88.5 percent. On the left you can see how active share rises as the result of stock picking. The fund doesn’t hold some of the stocks that are the index’s largest weighting and does hold stocks that are not in the index.

On the right is the active share as the result of sector weights (tracking error captures this well). Naturally, active share is a blend of stock picking and sector bets. These amounts are related. In general, high active share via stock picking and relatively low active share via sector bets does best.

Total Active Share Sector Active Share

The views expressed in this commentary reflect those of Legg Mason Capital Management (LMCM) as of the date of this commentary. These views are subject to change at any time based on market or other conditions, and LMCM disclaims any responsibility to update such views.

These views may not be relied upon as investment advice and, because investment decisions for clients of LMCM are based on numerous factors, may not be relied upon as an indication of trading intent on behalf of the firm. The information provided in this commentary should not be considered a recommendation by LMCM or any of its affiliates to purchase or sell any security. To the extent specific securities are mentioned in the commentary, they have been selected by the author on an objective basis to illustrate views expressed in the commentary. If specific securities are mentioned, they do not represent all of the securities purchased, sold or recommended for clients of LMCM and it should not be assumed that investments in such securities have been or will be profitable. There is no assurance that any security mentioned in the commentary has ever been, or will in the future be, recommended to clients of LMCM. Employees of LMCM and its affiliates may own securities referenced herein. Predictions are inherently limited and should not be relied upon as an indication of actual or future performance. Legg Mason Capital Management, LLC consists of two legal entities, Legg Mason Capital Management and LMM LLC.

Important risks All investments are subject to risks, including loss of principal. Small-cap stocks involve greater risks and volatility than large-cap stocks.

This document is for information only and does not constitute an invitation to the public to invest. You should be aware that the investment opportunities described should normally be regarded as longer term investments and they may not be suitable for everyone. The value of investments and the income from them can go down as well as up and investors may not get back the amounts originally invested, and can be affected by changes in interest rates, in exchange rates, general market conditions, political, social and economic developments and other variable factors. Past performance is no guide to future returns and may not be repeated. Investment involves risks including but not limited to, possible delays in payments and loss of income or capital. Neither Legg Mason nor any of its affiliates guarantees any rate of return or the return of capital invested. Please note that an investor cannot invest directly in an index. Forward-looking statements are subject to uncertainties that could cause actual developments and results to differ materially from the expectations expressed. This information has been prepared from sources believed reliable but the accuracy and completeness of the information cannot be guaranteed and is not a complete summary or statement of all available data. Individual securities mentioned are intended as examples of portfolio holdings and are not intended as buy or sell recommendations. Information and opinions expressed by either Legg Mason or its affiliates are current as of the date indicated, are subject to change without notice, and do not take into account the particular investment objectives, financial situation or needs of individual investors. The information in this document is confidential and proprietary and may not be used other than by the intended user. Neither Legg Mason nor any officer or employee of Legg Mason accepts any liability whatsoever for any loss arising from any use of this document or its contents. This document may not be reproduced, distributed or published without prior written permission from Legg Mason.

Distribution of this document may be restricted in certain jurisdictions. Any persons coming into possession of this document should seek advice for details of, and observe such restrictions (if any).

This document may have been prepared by an advisor or entity affiliated with an entity mentioned below through common control and ownership by Legg Mason, Inc.

This material is only for distribution in the jurisdictions listed.

Investors in Europe:

Issued and approved by Legg Mason Investments (Europe) Limited, registered office 201 Bishopsgate, London EC2M 3AB. Registered in England and Wales, Company No. 1732037. Authorized and regulated by the Financial Services Authority. Client Services +44 (0)207 070 7444. This document is for use by Professional Clients and Eligible Counterparties in EU and EEA countries. In Switzerland this document is only for use by Qualified Investors. It is not aimed at, or for use by, Retail Clients in any European jurisdictions.

Investors in Hong Kong, Korea, Taiwan and Singapore:

This document is provided by Legg Mason Asset Management Hong Kong Limited in Hong Kong and Korea, Legg Mason Asset Management Singapore Pte. Limited (Registration Number (UEN): 200007942R) in Singapore and Legg Mason Investments (Taiwan) Limited (Registration Number:

(98) Jin Guan Tou Gu Xin Zi Di 001; Address: Suite E, 55F, Taipei 101 Tower, 7, Xin Yi Road, Section 5, Taipei 110, Taiwan, R.O.C.; Tel: (886) 2-8722 1666) in Taiwan. Legg Mason Investments (Taiwan) Limited operates and manages its business independently. It is intended for distributors use only in respectively Hong Kong, Korea, Singapore and Taiwan. It is not intended for, nor should it be distributed to, any member of the public in Hong Kong, Korea, Singapore and Taiwan.

Investors in the Americas:

This document is provided by Legg Mason Investor Services LLC, a U.S. registered Broker-Dealer, which may include Legg Mason International - Americas Offshore. Legg Mason Investor Services, LLC, Member FINRA/SIPC, and all entities mentioned are subsidiaries of Legg Mason, Inc.

Investors in Canada:

This document is provided by Legg Mason Canada Inc. Address: 220 Bay Street, 4th Floor, Toronto, ON M5J 2W4. Legg Mason Canada Inc. is affiliated with the Legg Mason companies mentioned above through common control and ownership by Legg Mason, Inc.

Investors in Australia:

This document is issued by Legg Mason Asset Management Australia Limited (ABN 76 004 835 839, AFSL 204827) (“Legg Mason”). The contents are proprietary and confidential and intended solely for the use of Legg Mason and the clients or prospective clients to whom it has been delivered. It is not to be reproduced or distributed to any other person except to the client’s professional advisers.

This material is not for public distribution outside the United States of America.

Materials were prepared by Brandywine Global Investment Management, LLC and distributed by Legg Mason Investor Services, LLC.

Legg Mason Perspectives® is a registered trademark of Legg Mason Investor Services, LLC.

© 2012 Legg Mason Investor Services, LLC. Member FINRA, SIPC. Legg Mason Capital Management of Legg Mason, LLC. and Legg Mason Investor Services, LLC and all entities mentioned above are subsidiaries of Legg Mason, Inc. 406921 MIPX013439 3/12 FN1210649