Active managementmostlyefficientmarkets faj

Financial Analysts Journal
Volume 67  Number 6
©2011 CFA Institute

Active Management in Mostly Efficient Markets
Robert C. Jones, CFA, and Russ Wermers

This survey of the literature on the value of active management shows that the average active
manager does not outperform but that a significant minority of active managers do add value.
Further, studies suggest that investors may be able to identify superior active managers (SAMs)
in advance by using public information. Investors who can identify SAMs should be able to improve
their overall Sharpe ratio by including a meaningful exposure to active strategies.

M
ost debates have a clear winner: Lincoln Further, owing to data availability, most of the
beat Douglas, Kennedy beat Nixon, and surveyed studies analyzed only U.S. domestic
Reagan beat Mondale. But the debate equity mutual funds. Results for other asset classes
surrounding active versus passive man- (e.g., fixed income, non-U.S. equities, real estate,
agement continues to rage after more than 40 years commodities) and other active vehicles (e.g., sepa-
of contention. Should investors be satisfied with rate accounts, hedge funds, exchange-traded funds
index returns, or should they seek to outperform [ETFs]) are much more sparse or even nonexistent.
the indices by gathering and analyzing information The available data may contain survivor bias
that may already be reflected in security prices? (exclude dead funds or products that were closed
Advocates on both sides of the issue are dogmatic for poor performance) or self-selection bias
in their beliefs and provide compelling arguments (include only those managers who chose to submit
to support their cases. We surveyed the academic data) or have a limited return frequency (annually
research on the issue—covering both theory and or quarterly instead of monthly or daily). In addi-
empirical analysis—in order to offer investors tion, we note that the sample of actively managed
some practical advice. Our goal was to answer the ETFs is too small and too recently introduced for a
following three important questions for investors: statistically reliable analysis. Nevertheless, this
1. Does active management add value? concept is receiving a lot of attention in the industry
2. Can we identify superior active managers and should be a focus of future academic research.
ex ante? Moreover, we suspect that many of the results for
3. How much active risk should investors include active mutual funds would also apply to active
in their portfolios? ETFs because the latter are often managed by the
same managers using the same strategies.
Some Caveats We have tried to point out where our conclu-
The academic studies in this area are extensive, and sions likely apply—or do not apply—to these other
in some cases, the results are not entirely consistent. asset classes and vehicles. In general, however, we
The conclusions often depend on the period cov- should expect institutional funds (e.g., separate
ered, the methodology used, the universe and type accounts, pooled trust funds) to outperform retail
of funds considered, and the authors’ biases. funds (e.g., retail mutual funds) because the former
Although we have attempted to distill these varied (1) usually have lower management fees owing to
results into practical advice by giving more weight scale economies, (2) can use more performance-
to studies with more extensive data and more sensitive fees to better align the manager’s interests
robust analysis, there is always a chance that our with those of the investor, and (3) have lower costs
own biases may have colored our interpretation of for client accounting, client servicing, and manag-
the results or that the future will differ significantly ing daily cash flows.
from the periods covered in these studies. Finally, even where a study’s results are intui-
tive and statistically significant, they apply only on
Robert C. Jones, CFA, is chairman of Arwen Advisors, average, not necessarily to a specific manager or
Newark, New Jersey. Russ Wermers is associate profes- fund. For example, Chevalier and Ellison (1999)
sor of finance at the Robert H. Smith School of Business, found that managers who graduated from colleges
University of Maryland, College Park. whose students had higher average SAT scores

November/December 2011 www.cfapubs.org 29


outperformed other managers. This finding does (1997). Although we did not address after-tax
not mean, however, that all such managers will results in our survey, some studies (e.g., Dickson
outperform or that managers from other schools and Shoven 1993) have found that it is even harder
will not outperform. Thus, all our recommenda- for actively managed funds to outperform passive
tions are generic rather than specific to individual alternatives on an after-tax basis. In fact, the aver-
funds or managers. age underperformance reported in these studies is
only slightly lower in magnitude than the average
fee charged by active managers, which suggests
Does Active Management Add
that the average active manager earns a positive
Value? alpha before fees but that this alpha does not quite
If we assume that the aggregate of all actively man- cover the costs of active management. Further, this
aged funds is equal to the market—that is, active conclusion seems to apply equally to domestic
management is a zero-sum game—then the aggre- equity funds, fixed-income funds, international
gate of active fund returns equals the market return equity funds, balanced funds, and possibly even
but incurs trading costs and charges fees, and so the hedge funds and private equity vehicles (in the last
aggregate will underperform the market by an two cases, the data are not as readily available and
amount equal to fees and expenses. Empirical stud- the results are less compelling).2
ies seem to broadly support this conclusion.1 Proponents of active management often argue
Following Jensen’s seminal study (1968), that its benefits are most pronounced in periods of
numerous studies have reached virtually the same heightened volatility and economic stress.
conclusion: The average actively managed mutual Although no relevant academic studies yet exist for
fund does not capture alpha, net of fees and expenses. the period of the global financial crisis (2007–2009),
Figure 1 summarizes the results from two recent Standard & Poor’s has made some comparisons.3
studies (Barras, Scaillet, and Wermers 2010; Busse, The results for the five years ended December 2010
Goyal, and Wahal 2010). It shows that neither the show a fairly balanced “scorecard” for active versus
average mutual fund nor the average institutional passive management. Specifically, although the
separate account (ISA) earned a positive alpha, net market indices outperformed a majority of active
of fees and expenses, after adjusting for market and managers across all major international and domes-
style risks by using Carhart’s four-factor model tic equity categories, asset-weighted averages of

Figure 1. Four-Factor Alphas for Active Equity Mutual Funds and Active
Equity ISAs, Net of Fees and Expenses

Annualized Four-Factor Alpha (%)
0.3

0.2
0.10
0.1

0

–0.1
–0.12
–0.2

–0.3

–0.4

–0.5 –0.45 –0.47
–0.48
–0.53
–0.6
All Funds All Growth Aggressive Growth and Equal- Value-
Funds Growth Income Weighted Net Weighted Net

Based on Active Equity Mutual Fund Data (January 1975–December 2006)
Based on Active Equity Institutional Separate Account Data (1991–2007)

Sources: Barras, Scaillet, and Wermers (2010) for mutual fund data; Busse, Goyal, and Wahal (2010) for
ISA data.

30 www.cfapubs.org ©2011 CFA Institute


active managers matched or slightly beat the bench- In the Grossman–Stiglitz equilibrium (1980),
marks in most categories, with the exception of mid- the marginal (least skilled) active manager earns an
caps, international equities, and emerging markets. excess return that equals the costs of actively man-
These five-year results are somewhat worse for aging the assets (i.e., the costs of gathering and
actively managed bond funds. With the exception analyzing information, including the cost of human
of emerging-market debt, more than 50 percent of capital). Thus, in equilibrium, we should expect to
active managers failed to beat their benchmarks. see a (possibly large) group of marginal, yet active,
Although five-year asset-weighted average returns managers who just barely earn back their fees in the
were lower for active funds in all but three catego- form of excess returns. In the real world, however,
ries, equal-weighted returns over the same invest- the average active fund underperforms the index, net
ment horizon lagged in every category. of fees. Why is this so?
So, the most recent five years, which include Funds must provide other benefits—liquidity,
the global financial crisis, have apparently been a custody, bookkeeping, scale, optionality, or
bit more favorable to active equity management diversification—that justify their fees. In fact,
than the much longer periods covered by prior virtually 100 percent of passive funds underperform
studies—but not significantly so. As we will see their relevant indices, net of fees, which can average
later, some researchers have documented that anywhere from less than 15 bps a year for passive
active funds are more likely to outperform during mutual funds that invest in large-cap U.S. equities
financial downturns and recessions. Longer-term to more than 75 bps a year for passive emerging-
results that include both “stressed” and “normal” market funds.5 Thus, the relevant comparison is to
markets, however, offer little support for the “aver- the passive alternative and not to the index itself.
age” active manager. On that basis, the average active fund earns
roughly the same return as the average passive
fund, net of fees.6
Active Management and “Mostly The world we see around us—one filled with
Efficient Markets” active managers who, on average, provide no
excess return (versus the passive alternative), net
Although the semi-strong form of the efficient mar-
of fees—seems perfectly consistent with the
ket hypothesis—whereby market prices com-
Grossman–Stiglitz model (1980) of “mostly effi-
pletely and accurately reflect all publicly available cient” markets, wherein prices may not fully
information—predicts that rational investors reflect costly information. In the real world, with
(without inside information) will always choose differential skills among active managers, we
passive management over active management, this expect to (and do) find informed traders (or supe-
is clearly not the case. Why not? Are investors that rior active managers) who earn meaningful excess
irrational? If so, how can the market be efficient? In returns commensurate with their superior ability
their seminal paper, Grossman and Stiglitz (1980) to gather and analyze information. Of course, to
argued that in a world of costly information, the extent that they manage money for others, they
informed traders must earn an excess return or else are also likely to charge higher fees for their ser-
they would have no incentive to gather and analyze vices. Because of randomness in markets, how-
information to make prices more efficient (i.e., ever, discriminating perfectly between superior
reflective of information).4 In other words, markets and inferior active managers is nearly impossible
need to be “mostly but not completely efficient” or for investors to do ex ante; therefore, superior
else investors would not make the effort to assess active managers are unable to capture all their
whether prices are “fair.” If that were to happen, added value in the form of higher fees.7 So, the
prices would no longer properly reflect all available question becomes, Can we identify, ex ante, supe-
and relevant information and markets would lose rior active managers whose added value exceeds
their ability to allocate capital efficiently. Therefore, their fees? Several academic studies have offered
although the contest between active managers may encouraging results.
be a zero-sum game, active management as a whole
is definitely not: By making markets more efficient,
active management improves capital allocation—and
Identifying Superior Active
thus economic efficiency and growth—resulting in Managers
greater aggregate wealth for society as a whole. Thus, In a zero-sum game, if some investors (superior
we can view the excess returns earned by informed active managers, or SAMs) earn positive alphas,
traders as a kind of economic rent for gathering and other investors (inferior active managers, or IAMs)
processing information and thereby making mar- must “earn” negative alphas. SAMs can exploit
kets more efficient. IAMs in one of two ways: (1) by having superior



information and analytical capabilities, allowing percent to 60 percent. Figure 2, which is from their
them to better predict results, or (2) by providing paper, shows that the top-decile prior-alpha funds
liquidity/immediacy when IAMs need to trade produce annual future alphas of about 150 bps, net
quickly for reasons unrelated to expectations (e.g., of fees. Pástor and Stambaugh (2002) found that
to meet redemptions). Of course, the quality of an adjusting for sector biases can further improve
investor’s forecasts can vary across stocks and over these results.
time. In addition, today’s liquidity suppliers can be Also shown in Figure 2 are results from
tomorrow’s liquidity demanders; luck and ran- Kosowski, Timmermann, Wermers, and White
domness can also play an important role. Thus, (2006), who found that adjusting for non-normality
active management amounts to a SAM–IAM con- in fund alphas (e.g., skewness and fat tails)
test, in which telling exactly who is who can be improves the odds of identifying funds with
difficult (despite what Horton might hear)!8 greater return persistence. The intuition is that
Barras, Scaillet, and Wermers (2010) argued some SAMs may have positively skewed returns,
that some managers do have skills that allow them which are valuable to investors. Therefore, adjust-
to outperform the market, net of fees, but that ing for “random skewness” (defined as observed
distinguishing these skilled managers from man- fund return skewness that is statistically insignifi-
agers whose strong performance is simply due to cant) should help discriminate between managers
luck is virtually impossible. This argument may be who were lucky and those whose active returns
valid if investors have only simple past returns to exhibit a positive bias (or skew).
work with, but academics have discovered a num- Although research has shown that persistence
ber of ways for investors to identify SAMs by using appears to be strongest at short intervals (monthly
other data and methods. We recommend that or quarterly), some studies have documented per-
investors consider all four of the following factors sistence at intervals as long as three to four years.10
in identifying and evaluating SAMs: (1) past per- None of these studies, however, accounted for
formance (properly adjusted), (2) macroeconomic potential fund rebalancing costs, which would
forecasting, (3) fund/manager characteristics, and reduce the potential value added from picking
(4) analysis of fund holdings. active managers solely on the basis of prior returns
or risk-adjusted performance.
Past Performance. If SAMs do have superior Some researchers have also found evidence of
information or analytical capabilities, this advan- persistent skills among hedge fund managers.
tage should presumably persist over time. If so, Kosowski, Naik, and Teo (2007) extended the boot-
SAMs who outperform in one period will likely strap methodology of Kosowski, Timmermann,
outperform in the next period as well. Offsetting Wermers, and White (2006) into the hedge fund
this likelihood are various agency effects and com- universe. Accounting for the non-normal returns of
petitive pressures: Managers at poorly performing hedge funds is particularly important because they
funds are more likely to change their strategies or often pursue dynamic strategies and make security
be replaced, resulting in better future perfor- choices (e.g., various derivatives) that render their
mance.9 Managers at top-performing funds may portfolio returns distinctly non-normal. Kosowski,
lose their competitive edge—perhaps owing to Naik, and Teo (2007) found that top-decile hedge
changing technologies for security analysis, dimin- funds (ranked by adjusted two-year lagged Bayes-
ished motivation, or overconfidence—or they may ian posterior alpha) outperformed bottom-decile
reduce fund risk to protect their reputations and funds by 5.8 percent in the following year. Using a
fees. Equally problematic are top-performing man- simple ordinary least-squares ranking, they found
agers who may either raise fees or attract too many no significant difference between the top and bot-
assets for them to continue delivering the same net tom deciles, which illustrates the importance of
excess returns that they had in the past (i.e., at some adjusting hedge fund returns for non-normality.
point, there are diseconomies of scale in active man- Using the Fung–Hsieh seven-factor model
agement). Thus, lack of persistence does not neces- (2004), Fung, Hsieh, Naik, and Ramadorai (2008)
sarily imply a lack of skill among managers. investigated the performance, risk, and capital for-
Although the evidence is mixed, it seems to mation of funds of hedge funds over 1995–2004.
indicate that mutual fund returns exhibit modest Although the average fund of funds delivered alpha
persistence but only if excess returns are adjusted only over October 1998–March 2000, a subset of
to account for style biases by using either the Car- funds of funds delivered persistent alpha over lon-
hart four-factor model (1997) or the Fama–French ger periods. Such alpha-producing funds experi-
three-factor model (1993). Typical of the many ence steadier capital inflows and are less likely to
relevant studies, Harlow and Brown (2006) found liquidate. Those capital inflows reduce, but do not
that using style-adjusted returns can improve the eliminate, alpha producers’ ability to deliver alpha
odds of finding an outperforming fund from 45 in future periods.



Figure 2. Persistence in Past Performance

Decile by Prior Alpha

1

2

3

4

5

6

7

8

9

10

–4 –3 –2 –1 0 1 2
Future Annual Alpha (%)
Kosowski, Timmermann, Wermers, and White Four-Factor Alpha Model
Harlow and Brown Three-Factor Alpha Model

Notes: Harlow and Brown (2006) used a three-factor alpha methodology and rebalanced quarterly over
1979–2003. Using a four-factor alpha methodology with a three-year ranking period and a bootstrapping
technique to model non-normality, Kosowski, Timmermann, Wermers, and White (2006) rebalanced
annually over 1978–2002.
Sources: Harlow and Brown (2006); Kosowski, Timmermann, Wermers, and White (2006).

Although far fewer in number, studies of other The study by Busse, Goyal, and Wahal (2010)
asset classes and vehicles have provided findings was one of the few to address ISAs (see Figure 1).
that are generally consistent with those for hedge They found up to a year of weak persistence for
funds and domestic equity mutual funds. Kaplan domestic equity and fixed-income funds but less
and Schoar (2005) found persistence in the perfor- persistence for international equity funds. They did
mance of funds run by a private equity fund, not adjust for systematic risks (e.g., country, cur-
although (unlike our other investment categories) rency, style), however, which may explain the
general partners can sometimes polish perfor- weaker persistence for international fund returns
mance by using smoothed estimates for the “mar- in their study.
ket” values of their investments.11 Also unclear is Instead of examining persistence itself, Goyal
and Wahal (2008) looked at the hiring and firing
whether these general partners choose better
decisions of institutional sponsors and found that
investments or apply superior skills in overseeing
managers whom institutions hire do not outper-
the funded companies (or both).
form managers they fire.12 Because the hiring and
Bers and Madura (2000) found that actively
firing of managers involves search and transition
managed closed-end funds (CEFs) exhibit return costs,13 the implication is that institutions should
persistence. This finding is notable because CEFs not be so eager to change managers. But because
(like ISAs) do not need to deal with regular inflows institutions often change managers for reasons
and outflows, which can make detecting skills other than performance, this finding is not direct
harder (because managers of open-end funds evidence against persistence, though it does imply
must trade for both liquidity and informational that investors should be careful when changing
purposes). Huij and Derwall (2008) found evi- managers. That is, when making such decisions,
dence of persistence in fixed-income fund returns, investors should properly adjust past returns and
especially high-yield funds, after adjusting for consider other factors, including transition costs
multiple benchmarks. (additional factors are discussed later in the article).



When assessing past performance, distin- right conditions? The results seem quite promising,
guishing between the contribution of the manager but the strategies involve considerable manager
and the contribution of the fund management com- turnover and may not be so rewarding after
pany can be difficult; a SAM usually needs a strong deducting manager transition costs.
support team to perform well. Attempting to make To the extent that a fund’s alpha varies system-
that distinction, Baks (2003) found that, depending atically over time, this variation could be due to (1)
on the investment process, anywhere from 10 to 50 embedded macroeconomic sensitivities (e.g., a per-
percent of return persistence is due to the manager, sistent overweight in cyclical stocks), (2) time-
with the balance attributable to the fund manage- varying skills, or (3) time-varying opportunities for
ment company or other factors. Thus, investors managers to benefit from their skills. Although all
may want to discount prior performance following three explanations may play a role, studies lend the
a recent change in fund managers, especially for most support to the third one—namely, that cer-
funds that rely on “star” managers as opposed to a tain environments offer more mispricing opportu-
team approach or quant process. Similarly, inves- nities for managers to take advantage of their
tors should be cautious about chasing star manag- superior insights. For example, many contrarian
ers from one firm to another. managers underperformed during the tech bubble
Exhibit 1 summarizes several studies of persis- of the late 1990s, when prices diverged signifi-
tence. On the basis of these and other studies, we cantly from fundamentals, but outperformed by a
conclude that investors can likely identify SAMs by huge margin when the bubble burst. Did their
analyzing past performance, but they should be skills suddenly change dramatically, or did the
careful to account for non-normal return distribu- market simply provide more opportunities for
tions (such as those with skewness or fat tails) and them in one period versus the other?
various style/sector biases by using a sophisticated Moskowitz (2000) and Kosowski (2006) found
performance attribution system. Using less sophis- that the average active manager is more likely to
ticated techniques greatly diminishes one’s ability outperform the market during recessions than at
to find true SAMs on the basis of past performance.
other times. This outcome is probably not the result
In addition, investors should carefully assess
of holding cash in down markets because Kosowski,
whether any potential improvement from switch-
in particular, adjusted returns for market risk.
ing managers will cover transition costs.
Instead, recessions are likely to be periods of above-
Macroeconomic Forecasting. Studies of average uncertainty, when superior information
macroeconomic forecasting try to determine and analysis can be particularly valuable. Consis-
whether active managers, in general, perform bet- tent with this explanation, Kosowski, among others,
ter in certain environments and whether it is possi- also found that the average active fund performs
ble to predict, ex ante, which managers will perform better in periods of higher return dispersion and
best in a given environment. In other words, can volatility, which are also likely to be periods of
IAMs become SAMs (and vice versa) under the heightened uncertainty—and opportunity.

Exhibit 1. Summary of Findings on Persistence
Study Asset Class (period) Finding
Bers and Madura (2000) Active closed-end funds (1976–1996) Return
Harlow and Brown (2006) Equity mutual funds (1979–2003) Style-adjusted return
Huij and Derwall (2008) Fixed-income funds (1990–2003) Benchmark-adjusted return
Bollen and Busse (2005) Equity mutual funds (1985–1995) Raw return during quarter +1
Baks (2003) Equity mutual funds (1992–1999) Fund-manager-specific return
Kosowski, Timmermann, Wermers, and Equity mutual funds (1975–2002) Bootstrapping-derived alpha
White (2006)
Busse, Goyal, and Wahal (2010) Institutional separate accounts Weak return for domestic equity
(1991–2008)
Fung, Hsieh, Naik, and Ramadorai (2008) Hedge funds of funds (1995–2004) Alpha
Goyal and Wahal (2008) Institutional separate accounts None for termination
(1994–2003)
Jagannathan, Malakhov, and Novikov Hedge funds (1996–2005) Superior funds only
(2010)
Kosowski, Naik, and Teo (2007) Hedge funds (1994–2003) Alpha computed by using Bayesian methods
Kaplan and Schoar (2005) Private equity funds (1980–2001) Discounted cash inflows divided by outflows



Figure 3 shows the results from Kosowski produces annual four-factor alphas of more than
(2006); excess returns are higher in recessions than 600 bps, net of fund expenses but before any fund
in expansions for all the major Lipper domestic rebalancing costs. Banegas, Gillen, Timmermann,
equity fund categories. Figure 4, from a report by and Wermers (2009) applied the same approach
Paul (2009), shows that the top-quartile manager’s (with additional predictor variables) to European
alpha tends to be higher in periods of higher return equity funds and found even greater levels of out-
dispersion; in general, Paul found that the median performance (10–12 percent a year) owing to the
active U.S. domestic equity fund exhibits a quar- greater opportunities offered by rotating among
terly increase in alpha of 11–12 bps for every 1 country-specific mutual funds.
percent increase in the spread between the 25th and Avramov, Kosowski, Naik, and Teo (forth-
75th percentile stock returns in that quarter. coming 2011) extended the predictability models of
These studies focused on fund performance Avramov and Wermers (2006) into the hedge fund
during periods of economic stress. Such periods, universe and found a substantial ability to predict
however, are hard to predict. What if we want to hedge fund outperformance. Specifically, they
predict fund performance by using macrodata that found that several macroeconomic variables,
are known today? including the Chicago Board Options Exchange
One of the first studies of macroeconomic fore- Volatility Index (VIX) and credit spreads, allowed
casting to find superior asset managers was by them to identify hedge funds that outperformed
Avramov and Wermers (2006), who identified out- their Fung–Hsieh benchmarks (2004) by more than
performing funds, ex ante, by using macroeco- 17 percent a year, before transition costs (which can
nomic variables that have been shown to predict be substantial).
stock returns—namely, the level of short-term Note that a macroforecasting strategy can
rates, the credit default spread, the term structure often conflict with a return-persistence strategy for
of interest rates, and the market’s dividend yield. selecting funds—that is, macroforecasts may rec-
They found that selecting funds on the basis of ommend buying funds that have recently per-
their prior correlations with these macrovariables formed poorly, especially when macroeconomic

Figure 3. Alpha Performance during Recession and Expansion, 1962–2005

A. All Funds B. All Growth Funds C. Aggressive Growth
Annualized Four-Factor Alpha (%) Annualized Four-Factor Alpha (%) Annualized Four-Factor Alpha (%)
5 4 1
4 3
3
2 2 0
1 1
0 0
–1 –1 –1
–2
–3 –2
–4 –3 –2
Full Recession Expansion Full Recession Expansion Full Recession Expansion
Sample Sample Sample

D. Growth E. Growth and Income F. Balanced and Income
Annualized Four-Factor Alpha (%) Annualized Four-Factor Alpha (%) Annualized Four-Factor Alpha (%)
4 4 6
3 3 4
2 2 2
1 1
0
0 0
–1 –1 –2
–2 –2 –4
–3 –3 –6
Full Recession Expansion Full Recession Expansion Full Recession Expansion
Sample Sample Sample

Source: Kosowski (2006).



Figure 4. Manager Alpha and Return Dispersion, 1980–2007

Percent
35

30 Return Dispersion

25

20

15

10
Manager
Alpha
5

0

–5
80 89 98 07

Notes: Data are through 31 December 2007. Return dispersion is the difference in quarterly total returns between the 25th and 75th
percentiles of stocks in the AllianceBernstein U.S. large-cap universe. Manager alpha is the excess returns of the 25th percentile
managers in the eVestment U.S. large-cap equity universe versus the S&P 500 Index.
Source: Paul (2009).

conditions have recently changed. For example, if formance. See Exhibit 3 for a summary of findings
some funds have better relative returns when on fund/manager characteristics. Because fund
short-term interest rates are low and default management companies use different techniques
spreads are wide, we may decide to invest in such and are organized differently and because fund
funds when those conditions prevail, even if their managers come from a variety of backgrounds,
recent performance has been weak. certain fund companies and fund managers may be
A word of caution is in order: Macroeconomic more skilled than others at collecting and analyzing
timing strategies can involve considerable fund information. Many studies have found that certain
turnover from one period to the next (i.e., 200–300 types of funds, fund managers, and fund manage-
percent annually if implemented without con- ment companies reliably outperform, on average.
straints). This approach could prove challenging Further, because these approaches involve fairly
for large institutional investors if the target funds low levels of manager turnover, we view them as
have flow restrictions in place, especially for hedge being among the most effective ways to select
fund investors with lockup periods or penalties for active managers.
early withdrawal. Experienced managers are more likely than
Exhibit 2 summarizes the relevant studies of inexperienced managers to be SAMs because
macroeconomic timing. For investors who wish to unsuccessful fund managers (IAMs) are likely to
pursue this approach, we recommend investing in a drop out of the pool over time. Ding and Wermers
diverse group of funds that do well in different macroen- (2009) found exactly that: Experienced managers of
vironments and reallocating among them at the margin large funds (i.e., with assets under management
on the basis of macroforecasts. This cautious approach above the median) outperform less experienced
may provide less gross alpha but should incur managers by 92 bps a year. Interestingly, the oppo-
significantly lower manager search and transition site is often true for small funds. Ding and Wermers
costs, as well as provide a partial hedge against (2009) attributed their findings to entrenchment:
sudden changes in the macroenvironment.14 An experienced manager of a small fund has likely
been unsuccessful (which is why the fund is small)
Fund/Manager Characteristics. Various but may be difficult to replace for institutional or
researchers have studied the characteristics of other reasons. These findings point to a “mostly
funds, fund management companies, and fund efficient” market for fund managers, in which most
managers to see whether they can predict outper- (but not all) managers survive on the basis of skill.



Exhibit 2. Summary of Findings on Macroeconomic Timing
Study Finding
Moskowitz (2000) Average active managers outperform in recessions
Kosowski (2006) Average active managers outperform in periods of recession and high volatility/
dispersion
Avramov and Wermers (2006) Outperformance from identifying funds that do well in certain past environments and
choosing them on the basis of the current environment
Avramov, Kosowski, Naik, and Teo Same as Avramov and Wermers (2006) for hedge funds
(forthcoming 2011)
Banegas, Gillen, Timmermann, and Wermers Same as Avramov and Wermers (2006) for European equity funds
(2009)

Exhibit 3. Summary of Findings on Fund/Manager Characteristics
Study Finding
Chevalier and Ellison (1999) Managers from schools with higher average SATs do better
Edelen (1999) Funds with low cash balances outperform
Liang (1999) Hedge funds (especially those with higher incentive fees and lockups) outperform
Howell (2001) Younger hedge funds do better
Wermers (2010) Funds with the most variation in style exposures do better
De Souza and Gokcan (2003) Managers who invest in their own funds outperform; hedge funds outperform
mutual funds; older, larger hedge funds with higher incentive fees outperform
other hedge funds
Amenc, Curtis, and Martellini (2004) Larger, younger hedge funds with high incentive fees outperform
Getmansky (2005) Hedge funds that are not too large or too small do better
Kacperczyk, Sialm, and Zheng (2005) Industry- or sector-concentrated funds do better
Gottesman and Morey (2006) Managers from better MBA programs outperform
Ding and Wermers (2009) Experienced managers of large funds outperform; the opposite is true for small funds
Cohen, Frazzini, and Malloy (2008) Stocks with direct school ties between board members and fund managers outperform
Massa and Zhang (2009) Funds with flatter organizational structures outperform
Dincer, Gregory-Allen, and Shawky (2010) Well-trained managers outperform
Kinnel (2010) Mutual fund managers with lower expense ratios outperform

Other manager characteristics that can help annual BusinessWeek rankings. Interestingly,
predict outperformance include social connections, they found no relationship between perfor-
academic background, and co-investment: mance and other graduate degrees (including
• Cohen, Frazzini, and Malloy (2008) found that the PhD degree) or the CFA designation.
managers take larger positions in companies in • Dincer, Gregory-Allen, and Shawky (2010)
which they have social connections (i.e., the found that funds managed by CFA charter-
officers or board members attended the same holders tend to have less tracking risk (better
college as the manager). Further, these hold- risk management) than other funds. Con-
ings outperform nonconnected holdings, on versely, funds managed by MBA graduates
average. They conjectured that connected (without the CFA designation) tend to have
managers have better access to private infor-
higher tracking risk, which may reflect their
mation or are better able to assess the quality
proverbial overconfidence. They found no sig-
of the company’s management team.
nificant difference in returns (as opposed to
• Chevalier and Ellison (1999) documented that
risk) attributable to either the MBA degree or
managers who graduate from colleges whose
students have higher average SAT scores also the CFA designation—or, for that matter, to
tend to outperform, presumably because they experience. Unlike Ding and Wermers (2009),
are better qualified and thus better able to however, they did not adjust for the correlation
analyze information.15 between experience and fund size.
• Similarly, Gottesman and Morey (2006) found • De Souza and Gokcan (2003) found that hedge
that the quality of a manager’s MBA program fund managers who invest their own capital
is positively correlated with future perfor- in their funds are more likely to outperform,
mance. They measured the quality of an MBA possibly because such managers have greater
program by using both the average GMAT conviction or are more likely to avoid uncom-
score of students in the program and the pensated risks.



On the basis of these studies, we conclude that past performance, however, we suspect that they
investors should look for funds directed by smart, would have less predictive ability than a more com-
well-networked, and well-educated managers who prehensive rating methodology.
have some skin in the game. Conflicting results surround the issue of spe-
Other studies have focused on the characteris- cialization versus flexibility. On the one hand, we
tics of the fund management company. Not surpris- would expect managers who stick to their areas of
ingly, funds sponsored by large management expertise (i.e., where they have a competitive advan-
companies tend to perform better than those spon- tage) to outperform those who do not. On the other
sored by small companies because of (1) economies hand, any constraint can limit a manager’s ability to
of scale and scope (lower costs and fees),16 (2) add alpha. The evidence supports both “hands.”
greater resources for gathering and analyzing Kacperczyk, Sialm, and Zheng (2005) found
information,17 and (3) better technologies for exe- that funds concentrated in certain industries and
cuting trades with less price impact. In addition, sectors outperform an appropriate benchmark,
fund management companies with a greater num- which suggests that specialization along industry
or sector lines improves a manager’s ability to
ber of independent directors also tend to perform
gather and analyze information. In addition, some
better, possibly because they are more demanding
of the macrobased studies (discussed previously)
of their managers.18
found that different types of managers and funds
Massa and Zhang (2009) found that funds man- do better in different environments, in which there
aged by companies with a flat organizational struc- are presumably more mispricing opportunities in
ture outperform funds managed by companies their primary areas of expertise.
with a more hierarchical structure. Funds managed Conversely, Wermers (2010) found that funds
by companies with a flat organizational structure that allow the most “style drift” (variation in expo-
also tend to be more concentrated and exhibit less sures to such style factors as value versus growth)
herding behavior. Massa and Zhang (2009) found are more likely than others to outperform their
that each additional layer in the hierarchy reduces benchmarks. In addition, many studies have
average fund performance by 24 bps a month, or shown that eliminating the “no short” constraint
almost 300 bps a year. The reason for this result may can improve a fund’s active risk–reward profile.19
be that a hierarchical structure discourages manag- Moreover, Liang (1999) found that hedge funds—
ers from innovating, taking risks, and collecting which usually have fewer constraints—exhibit
private information—that is, in vertical organiza- more skill and have higher risk-adjusted outperfor-
tions, managers may not think they have as much mance than mutual funds. We conclude that inves-
direct responsibility (ownership) for their funds. tors should look for funds with few artificial
Yet other studies have focused on the charac- constraints but whose managers do not stray too
teristics of the fund itself. Edelen (1999) found that far from their primary areas of expertise.
cash holdings explain much of the underperfor- Finally, a number of studies have examined the
mance of the average mutual fund. Thus, funds characteristics of top-performing hedge funds:
with large, idle cash balances are more likely to lag • Liang (1999) found that hedge funds with
their benchmarks than are other funds. If so, funds “high watermarks” significantly outperform
that equitize idle cash balances can eliminate this funds without such structures.20 He also found
that funds with higher incentive fees and lon-
“cash drag” as a source of underperformance. Con-
ger lockup periods perform better than other
versely, if cash is used as a tactical (market-timing)
funds. These results suggest that the best per-
tool, it may be a source of alpha. Careful attribution
forming funds are those that best align the
analysis should help investors discern whether
interests of the manager and the investor.
market timing is an alpha source or whether the
• Numerous studies have found that large hedge
manager should be equitizing idle cash balances. funds perform better than small hedge funds,
In a well-documented and remarkably objec- possibly because of economies of scale (see,
tive study from Morningstar, Kinnel (2010) found e.g., Amenc, Curtis, and Martellini 2004; Get-
that expense ratios (fees) are strong predictors of mansky 2005; De Souza and Gokcan 2003). For
performance: The cheapest funds outperformed example, larger funds may be able to attract
the most expensive funds in every period and every and retain more and better analysts and man-
category. Whether Morningstar’s star ratings had agers because of their higher revenue base.
any additional predictive ability after accounting Getmansky (2005), however, found a concave
for the higher net returns of low-fee funds was relationship between fund size and perfor-
unclear. Given that its star ratings rely solely on mance: Funds that are either too small or too



large underperform, which suggests an opti- month (or 216 bps a year), whereas those with a
mal size for most hedge funds. Funds that are large positive return gap outperform by about 10
too large may run up against liquidity con- bps a month (120 bps annually).
straints or possibly lose their competitive edge Huang, Sialm, and Zhang (2010) compared the
owing to wealth effects (i.e., a wealthy man- realized volatility of a fund (based on reported
ager with a solid reputation and track record returns) with the volatility calculated by using its
may have little incentive to aggressively seek most recently reported holdings. They found that
strong performance). “risk-shifting” funds tend to underperform funds
• Results are conflicting with respect to fund age. that maintain a stable risk profile. Fund managers
Howell (2001) found that younger funds (less who cut risk may be “locking in” gains to protect
than three years since inception) outperform their fees, whereas managers who increase risk
older funds by more than 700 bps a year, but this may be “doubling down” in a desperate attempt to
finding could be due to self-selection and back- catch up to other funds. In any case, Huang, Sialm,
fill bias (after the fact, only successful young and Zhang (2010) concluded that risk shifting
funds choose to submit information on their either is an indication of inferior ability or is moti-
returns to databases). Conversely, De Souza and vated by agency issues. We conclude that investors
Gokcan (2003) found that older funds outper- should select funds that manage risk effectively
form younger funds, on average. Given these and that maintain a reasonably stable (but not nec-
conflicting results, we would ignore fund age essarily low) risk profile.
and focus more on manager experience (and Managers who have superior information or
other factors) when selecting hedge funds. analytical capabilities are likely to be contrarians—
that is, because prices reflect consensus expecta-
Portfolio Holdings Analysis. Studies in this tions, SAMs will trade only when their views differ
area have looked at the holdings of the underlying from the prevailing consensus. Wei, Wermers, and
fund to determine whether there is any information Yao (2009) found exactly that: Contrarian managers
that can help predict performance. Holdings-based outperform herding managers21 by more than 260
analysis generally requires much more detailed bps a year. Their findings suggest that these excess
(holdings) data and involves additional computa- returns come both from supplying liquidity to the
tional complexity. Results seem quite promising, herd and from superior information collection and
however, and would appear to justify the additional analysis—in particular, the stock holdings of con-
analysis. In fact, because these approaches involve trarian funds show a much greater improvement in
low manager turnover and get at the heart of a future corporate profitability than do the holdings
manager’s strategy, we view holdings-based analy- of herding funds.
sis as one of the best ways to identify SAMs. See Cremers and Petajisto (2009) found that man-
Exhibit 4 for a summary of findings on holdings- agers who take big active positions perform better
based analysis. than those who take small positions. They defined
Kacperczyk, Sialm, and Zheng (2008) com- “active share” as the absolute difference between a
pared the performance of the actual fund with the stock’s weight in the portfolio and its weight in the
performance of the publicly disclosed holdings of “best fit” benchmark, cumulated across all the
the fund (the “return gap”). A large negative stocks in the portfolio and the benchmark. They
return gap may indicate sloppy trading or a man- found that funds with the highest aggregate active
ager who is trying to hide bad trades (i.e., “win- share outperform those with the lowest active share
dow dressing” at quarter-end, when holdings are by roughly 250 bps a year. They attributed this
published). Both explanations are cause for con- result to greater “conviction” on the part of the
cern. They found that funds with a large negative manager and concluded that “the most active stock
return gap underperform by roughly 18 bps a pickers have enough skill to outperform their

Exhibit 4. Summary of Findings on Holdings-Based Analysis
Study Finding
Cremers and Petajisto (2009) Funds that take bigger active positions outperform
Kacperczyk, Sialm, and Zheng (2008) Funds with large “return gaps” between published and
holdings-derived performance underperform
Huang, Sialm, and Zhang (2010) Funds that change risk underperform
Wei, Wermers, and Yao (2009) Contrarian managers (by holdings) outperform



benchmarks even after fees and transaction costs. will be zero or negative. Such investors should
In contrast, funds focusing on factor bets seem to take little or no active risk, regardless of volatilities
have zero-to-negative skill, which leads to particu- and correlations. They should focus on developing
larly bad performance after fees” (p. 3332). an appropriate asset allocation while embracing
There are some important qualifications, how- passive management in each asset class, secure in
ever, to the findings of Cremers and Petajisto the knowledge that by minimizing fees and
(2009). First, they did not control for the capitaliza- expenses, they are likely to perform at least as well
tion of the benchmark. Thus, their results could as the average investor in each asset class. Such
simply mean that small-cap funds outperform their investors are essentially piggybacking on “the wis-
benchmarks more often than large-cap funds do.22 dom of crowds.”24
Such a finding, though interesting, is entirely con- Nevertheless, many of the studies surveyed in
sistent with the “costly information” theory of this article offer reliable ways to identify SAMs on
Grossman and Stiglitz (1980) because small-cap an ex ante basis. Investors who can do so success-
stocks are more costly to research. Second, strong fully should be rewarded with positive expected
manager conviction is not always a good thing. The returns from active management. Further, active
behavioral literature shows that overconfidence risk, almost by definition, should have little corre-
can cause investors to take excessive risk relative to lation with the systematic risks in an investor’s
the quality of their information or analysis. There-
portfolio. Similarly, little correlation should exist
fore, a large active share may indicate a BUM
between the active risks of different active manag-
(“behaviorally unhumble manager”) rather than a
ers who presumably use different strategies.25
true SAM. Finally, Cremers and Petajisto’s own
analysis does not indicate any return differences Taken together, these observations imply that
between high and low tracking error (as opposed to active management can provide both positive
active share) funds, which implies that the best (active) returns and substantial diversification ben-
funds have high active shares but relatively low efits. For investors who can identify SAMs, a thor-
tracking errors (if the active returns are equivalent, ough risk budgeting exercise would likely
the lower tracking error funds will have higher recommend a healthy allocation to active strategies.
information ratios, ipso facto). Table 1 shows the potential benefits of adding
Combining these observations, we conclude active risk to a passive portfolio. Let us assume that
that investors should look for funds with a contrar- an investor starts with a passive portfolio that has
ian bent and high active shares but stable tracking an expected return of 7 percent and volatility of 10
errors. Managers of such funds are likely to be percent. Further, let us assume that the investor has
adept at risk management (i.e., at taking large but been able to identify a group of SAMs—which may
offsetting bets) but are probably not overconfident include hedge funds, active long-only funds, global
because they hedge their largest bets. Investors tactical asset allocation overlay managers, and so
should also avoid “risk shifters” and funds with a on—with an aggregate expected information ratio
large negative return gap. of 0.25 and no correlation with other risks in the
investor’s portfolio. If the investor decides to add 6
Active Risk Budgeting percent in active risk to the existing portfolio—
Risk budgeting is the process of allocating risk to without any reduction in other risks—the portfo-
different investment alternatives. Active risk bud- lio’s total expected return improves from 7 percent
geting means determining how much risk to allo- to 8.5 percent, whereas the expected volatility
cate to different active strategies—that is, how much increases from 10 percent to 11.7 percent.26 Thus,
and what type of active risk should investors the incremental return–risk ratio (IRRR) is 0.88, or
include in their portfolios?23 The answer is a func- 1.5 percent/1.7 percent. An investor with a 20-year
tion of the expected return to active risk (i.e., the horizon can realize a 32 percent increase in expected
expected information ratio) and the correlation
between that active risk and other risks in an inves-
tor’s portfolio. In essence, active risk budgeting is Table 1. Adding Active Risk to a Passive
an optimization exercise that requires estimates of Portfolio
the expected returns, volatilities, and correlations Passive Active Total Incremental
of different active strategies. Portfolio Strategies Portfolio Analysis
For investors who are unwilling or unable to Return 7.0% 1.5% 8.5% 1.5%
devote the level of resources necessary to identify Risk 10.0% 6.0% 11.7% 1.7%
SAMs, the expected return to active management IRRR 0.70 0.25 0.73 0.88



terminal wealth—often, enough to convert an they should focus on setting the right asset alloca-
underfunded plan into a fully funded one. Of tion and not worry too much about finding SAMs.
course, the slightly higher volatility also means a In fact, however, the BHB results say only that
wider range of expected terminal wealth. returns to asset classes explain the vast majority of
Figure 5, from Litterman (2004), shows the a typical fund’s return variance over time; they say
optimal allocation to active risk for a typical equity nothing about cross-sectional differences in actual
investor as a function of the expected information long-term returns.27
ratio (IR). For example, at an expected IR of 0.25, In recent years (i.e., after the 1986 BHB study),
a typical investor should take more than 6 percent we have seen some investors—primarily individu-
in active risk, far more than we see in most port- als, foundations, and endowments—take on signif-
folios. Litterman called this scenario “the active icantly more active risk, with a substantial impact
risk puzzle.” Why do we see the vast majority of on performance. In fact, many of the most successful
investors taking active risk of 2 percent or less— investors over the past decade (even after the recent
which implies an expected IR of only about 0.05— crisis) have had substantial exposures to active
instead of a bimodal distribution whereby many strategies—including hedge funds, active equity,
investors take no active risk and the remaining global tactical asset allocation overlays, currency
investors take substantially more active risk? Lit- overlays, private equity, commodities, alternative
terman suspected that agency issues may explain assets, and other sources of active risk. Clearly,
this anomaly—for example, institutional sponsors some of these investors have been able to identify
may seek to limit career risk by herding with their SAMs—as evidenced by their positive active
peers and being overcautious. returns—and have done quite well as a result.
Another possible explanation, however, may Finally, for investors with limited resources,
be that investors misinterpret the oft-cited study by focusing their search for SAMs on those asset
Brinson, Hood, and Beebower (BHB 1986), who classes whose rewards to active management are
found that asset allocation explains more than 90 likely to be greatest may make sense—that is, asset
percent of a typical plan’s return variance over classes for which information gathering and analy-
time. Investors may misinterpret this finding to sis is most difficult and expensive. For an investor to
mean that active management within asset classes embrace passive management in some asset classes and
has little impact on relative performance; therefore, active management in others is perfectly rational.

Figure 5. Optimal Active Risk Allocation for a Typical Equity Investor as a
Function of the Expected IR

Optimal Allocation to Active Risk (%)
8

6

4

2

0
0 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Aggregate Active Risk Information Ratio

Sources: Goldman Sachs Asset Management; Litterman (2004).



Conclusion Investors can lessen this risk by using some of
the research we have discussed. In particular,
Our review of academic studies of active manage-
studies suggest that investors may be able to iden-
ment has produced the following findings and
tify SAMs ex ante by considering (1) past perfor-
recommendations.
mance (properly adjusted), (2) macroeconomic
Active returns (adjusted for risk) across man-
correlations, (3) fund/manager characteristics,
agers and time probably average close to zero, net
and (4) analyses of fund holdings. We suspect that
of fees and other expenses. This finding is what we
using a combination of these approaches will pro-
should expect in a mostly efficient market, in
duce better results than following any one
which fierce competition among active managers
approach exclusively.
drives average (net) active risk-adjusted returns
Active management will always have a place in
toward zero, in equilibrium. By keeping markets
“mostly efficient” markets. Hence, investors who
efficient, however, active management provides a
can identify SAMs should always expect to earn a
critical function in modern capitalist economies:
relative return advantage. Further, this alpha can
Efficient, rational capital allocation improves eco-
have a substantial impact on returns with only a
nomic growth and leads to increased wealth for
modest impact on total portfolio risk. Finding such
society as a whole.
managers is not easy or simple—it requires going
Thus, to keep the competition fierce, the
well beyond assessing past returns—but academic
rewards to superior (as opposed to average) active
studies indicate that it can be done.
management must be rich indeed, as in fact they
are—for both the manager and the ultimate inves-
tor. Superior managers earn high fees and often A number of colleagues have contributed to this article.
share in their added value, whereas inferior man- We would particularly like to thank Andrew Alford,
agers are soon bereft of both clients and fees. Inves- Wesley Chan, Gary Chropuvka, Kent Clark, Kent
tors who engage active managers can earn positive Daniel, Terence Lim, Surbhi Mehta, and Sam Shikiar
alphas with modest additional risk on a total port- for their valuable insights and comments. The quality
of the article benefited greatly from their help.
folio basis (i.e., an attractive IRRR). But this benefit
comes at a cost—the risk that active returns may
This article qualifies for 1 CE credit.
prove negative and lead to lower terminal wealth.

Notes
1. French (2008) made this assumption in computing the 6. For instance, Wermers (2000) found that the asset-weighted
“deadweight loss” from investing in active funds and con- average actively managed domestic equity fund exhibited
cluded that 67 bps a year is lost from active fund invest- a net return over 1977–1994 that was the same as that of the
ments. Wermers and Yao (2010), however, found large Vanguard 500 Index Fund.
differences in aggregate holdings of active versus passive 7. Berk and Green (2004) presented a theoretical model that
funds, which suggests that this assumption is questionable. captures this intuition.
For instance, passive funds hold more large-cap core stocks,
8. With apologies to Dr. Seuss.
whereas active funds hold more small- and mid-cap stocks.
Nevertheless, several studies have found that, for whatever 9. See Lynch and Musto (2003) for a formal model with these
reason, the average active fund underperforms the market predictions.
by roughly the magnitude of fees and expenses. 10. With respect to short-term persistence, Bollen and Busse
2. For example, Ferson, Henry, and Kisgen (2006) evaluated a (2005) used daily mutual fund returns to rank actively
sample of U.S. government bond funds and found underper- managed U.S. domestic equity mutual funds on a quarterly
formance, net of fees; Phalippou and Gottschalg (2009) found basis. They found an average abnormal return of 39 bps (156
that private equity funds also underperform, net of fees. bps annually) for the top decile in the postranking quarter.
3. See year-end 2010 Standard & Poor’s Indices versus Active For longer-term persistence, Wermers (2003) found that
(SPIVA) Funds Scorecards (www.standardandpoors.com/ portfolios of top-decile funds, ranked by lagged one-year
indices/spiva/en/us). net return, outperform for three to four years.
4. Informed traders may have better information and/or a 11. To overcome this potential bias, Kaplan and Schoar (2005)
better ability to assess the implications of information for computed returns on the basis of discounted private equity
security prices. We use the terms informed trader and superior fund cash flows rather than potentially biased valuations
active manager interchangeably throughout this article. issued by the fund managers. Nevertheless, the lack of
5. Under both passive and active management, fees are lower reliable market pricing makes such studies less reliable.
for institutional products, but the point is the same: Active Also, the dataset used in the study was based on voluntary
managers should be compared with the passive alternative, reporting of fund returns by the private equity firms (or
not the index itself. general partners), as well as their limited partners.



12. Stewart, Neumann, Knittel, and Heisler (2009) also studied Lim, and Stein (2000) found the same for momentum. Chor-
the hiring and firing decisions of institutions. As they dia, Subrahmanyam, and Tong (2010) found that the value
noted, institutional plan sponsors are charged with invest- anomaly has been stronger on the short side in recent years.
ing more than $10 trillion in assets for pension plans, 20. With a high watermark, the fund does not earn a perfor-
endowments, and foundations. Using a dataset covering mance fee until it makes up any prior underperformance.
80,000 yearly observations of institutional investment 21. A herding manager trades (i.e., has a change in holdings) in
product assets, accounts, and returns over 1984–2007, they the same direction as the aggregate of other managers. A
found little evidence of value added by managers picked contrarian manager trades in the opposite direction from
by the sponsors over time. In fact, they estimated losses of the aggregate.
more than $170 billion over the period owing to poor
22. Small-cap funds are likely to have a larger active share
manager selection by sponsors.
than large-cap funds because the small-cap benchmark has
13. Transition costs include the costs of identifying and inter- more names, with a smaller average weight, than the large-
viewing new managers and the costs of converting portfo- cap benchmark.
lios to a new strategy/manager (trading costs). 23. For a thorough discussion of active risk budgeting, see
14. Although the strategies of Avramov and Wermers (2006) Winkelmann (2003).
and Avramov, Kosowski, Naik, and Teo (forthcoming 2011) 24. For a cogent explanation of why consensus views are often
hedge against some potential changes in the macroeco- more accurate than expert opinions, see Surowiecki (2004).
nomic environment, a more cautious strategy would pro-
25. A more formal risk budgeting exercise would include explicit
vide greater safety in the face of large, unexpected shifts.
estimates of these correlations and related volatilities.
15. Unable to find data on individual manager SAT scores,
26. This example assumes that the investor can add active risk
Chevalier and Ellison (1999) used college average SAT
without reducing systematic risk or return. Under this
scores as a proxy.
assumption, total return is 8.5 percent (7 percent + 0.25  6
16. Chen, Hong, Huang, and Kubik (2004) found economies of percent). Assuming no correlation between active risk and
scale in large management companies. other portfolio risks, the total risk is the square root of (10
17. Busse, Goyal, and Wahal (2010) found that funds spon- percent squared + 6 percent squared), or the square root of
sored by companies with broad research capabilities tend 136 percent = 11.7 percent.
to outperform. 27. For a more complete discussion, see Xiong, Ibbotson,
18. Ding and Wermers (2009) found that one additional inde- Idzorek, and Chen (2010), who demonstrated that the
pendent director for a fund is correlated with an additional results of BHB (1986) really mean that market movements
20 bps a year in pre-expense returns. explain most of a typical plan’s returns and that a plan’s
19. For example, Hirshleifer, Teoh, and Yu (2011) found that specific asset allocation and active management contribute
the accrual anomaly is stronger on the short side; Hong, similar amounts to total active return and risk.

References
Amenc, Noël, Susan Curtis, and Lionel Martellini. 2004. “The Bollen, Nicolas P.B., and Jeffrey A. Busse. 2005. “Short-Term
Alpha and the Omega of Hedge Funds Performance.” Unpub- Persistence in Mutual Fund Performance.” Review of Financial
lished paper, EDHEC (February). Studies, vol. 18, no. 2 (Summer):569–597.
Avramov, Doron, and Russ Wermers. 2006. “Investing in Brinson, Gary P., L. Randolph Hood, and Gilbert L. Beebower.
Mutual Funds When Returns Are Predictable.” Journal of Finan- 1986. “Determinants of Portfolio Performance.” Financial Ana-
cial Economics, vol. 81, no. 2 (August):339–377. lysts Journal, vol. 42, no. 4 (July/August):39–48. Reprinted in
Avramov, Doron, Robert Kosowski, Narayan Y. Naik, and Financial Analysts Journal, vol. 51, no. 1 (January/February 1995,
Melvyn Teo. Forthcoming 2011. “Hedge Funds, Managerial Skill, 50th Anniversary Issue):133–138.
and Macroeconomic Variables.” Journal of Financial Economics. Busse, Jeffrey A., Amit Goyal, and Sunil Wahal. 2010. “Perfor-
Baks, Klaas. 2003. “On the Performance of Mutual Fund Man- mance and Persistence in Institutional Investment Manage-
agers.” Unpublished paper, Emory University (March). ment.” Journal of Finance, vol. 65, no. 2 (April):765–790.

Banegas, Ayelen, Benjamin Gillen, Allen G. Timmermann, and Carhart, Mark M. 1997. “On Persistence in Mutual Fund Perfor-
Russ Wermers. 2009. “The Performance of European Equity mance.” Journal of Finance, vol. 52, no. 1 (March):57–82.
Mutual Funds.” Working paper, University of California, San Chen, Joseph, Harrison Hong, Ming Huang, and Jeffrey Kubik.
Diego, and University of Maryland (March). 2004. “Does Fund Size Erode Mutual Fund Performance? The
Barras, Laurent, Olivier Scaillet, and Russ Wermers. 2010. Role of Liquidity and Organization.” American Economic Review,
“False Discoveries in Mutual Fund Performance: Measuring vol. 94, no. 5 (December):1276–1302.
Luck in Estimated Alphas.” Journal of Finance, vol. 65, no. 1 Chevalier, Judith, and Glenn Ellison. 1999. “Are Some Mutual
(February):179–216. Fund Managers Better than Others? Cross-Sectional Patterns in
Berk, Jonathan, and Richard Green. 2004. “Mutual Fund Flows Behavior and Performance.” Journal of Finance, vol. 54, no. 3
and Performance in Rational Markets.” Journal of Political Econ- (June):875–899.
omy, vol. 112, no. 6 (December):1269–1295. Chordia, Tarun, Avanidhar Subrahmanyam, and Qing Tong.
Bers, Martina, and Jeff Madura. 2000. “Why Does Performance 2010. “Trends in the Cross-Section of Expected Stock Returns.”
Persistence Vary among Closed-End Funds?” Journal of Financial Unpublished paper, Emory University, UCLA, and Singapore
Services Research, vol. 17, no. 2 (August):127–147. Management University (November).



Cohen, Lauren, Andrea Frazzini, and Christopher Malloy. Hong, Harrison, Terence Lim, and Jeremy Stein. 2000. “Bad
2008. “The Small World of Investing: Board Connections and News Travels Slowly: Size, Analyst Coverage, and the Profit-
Mutual Fund Returns.” Journal of Political Economy, vol. 116, ability of Momentum Strategies.” Journal of Finance, vol. 55,
no. 5 (October):951–979. no. 1 (February):265–295.
Cremers, Martijn, and Antti Petajisto. 2009. “How Active Is Your Howell, Michael J. 2001. “Fund Age and Performance.” Journal
Fund Manager? A New Measure That Predicts Performance.” of Alternative Investments, vol. 4, no. 2 (Fall):57–60.
Review of Financial Studies, vol. 22, no. 9 (September):3329–3365. Huang, Jennifer, Clemens Sialm, and Hanjiang Zhang. 2010.
De Souza, Clifford, and Suleyman Gokcan. 2003. “How Some “Risk Shifting and Mutual Fund Performance.” Working paper,
Hedge Fund Characteristics Impact Performance.” AIMA Jour- University of Texas at Austin (May).
nal (September). Huij, Joop, and Jeroen Derwall. 2008. “Hot Hands in Bond
Dickson, Joel, and John Shoven. 1993. “Ranking Mutual Funds Funds.” Journal of Banking & Finance, vol. 32, no. 4 (April):559–572.
on an After-Tax Basis.” NBER Working Paper 4393 (July). Jagannathan, Ravi, Alexey Malakhov, and Dmitry Novikov.
Dincer, Oguzhan, Russell B. Gregory-Allen, and Hany A. 2010. “Do Hot Hands Exist among Hedge Fund Managers? An
Shawky. 2010. “Are You Smarter than a CFA’er?” Working Empirical Evaluation.” Journal of Finance, vol. 65, no. 1 (February):
paper, Illinois State University, Massey University, and SUNY 217–255.
at Albany (January). Jensen, Michael C. 1968. “The Performance of Mutual Funds in
Ding, Bill, and Russ Wermers. 2009. “Mutual Fund Performance the Period 1945–1964.” Journal of Finance, vol. 23, no. 2
and Governance Structure: The Role of Portfolio Managers and (May):389–416.
Boards of Directors.” Working paper, SUNY at Albany and Kacperczyk, Marcin, Clemens Sialm, and Lu Zheng. 2005. “On
University of Maryland (November). the Industry Concentration of Actively Managed Equity Mutual
Edelen, Roger M. 1999. “Investor Flows and the Assessed Funds.” Journal of Finance, vol. 60, no. 4 (August):1983–2011.
Performance of Open-End Mutual Funds.” Journal of Financial ———. 2008. “Unobserved Actions of Mutual Funds.” Review of
Economics, vol. 53, no. 3 (September):439–466. Financial Studies, vol. 21, no. 6 (June):2379–2416.
Fama, Eugene F., and Kenneth R. French. 1993. “Common Risk Kaplan, Steven N., and Antoinette Schoar. 2005. “Private Equity
Factors in the Returns on Stocks and Bonds.” Journal of Financial Performance: Returns, Persistence, and Capital Flows.” Journal
Economics, vol. 33, no. 1 (February):3–56. of Finance, vol. 60, no. 4 (August):1791–1823.
Ferson, Wayne, Terry R. Henry, and Darren J. Kisgen. 2006. Kinnel, Russell. 2010. “How Expense Ratios and Star Ratings
“Evaluating Government Bond Fund Performance with Sto- Predict Success.” Morningstar Advisor (10 August).
chastic Discount Factors.” Review of Financial Studies, vol. 19, Kosowski, Robert. 2006. “Do Mutual Funds Perform When It
no. 2 (Summer):423–455. Matters Most to Investors? U.S. Mutual Fund Performance and
French, Kenneth. 2008. “Presidential Address: The Costs of Risk in Recessions and Expansions.” Working paper, Imperial
Active Investing.” Journal of Finance, vol. 63, no. 4 (August): College Business School (August).
1537–1573. Kosowski, Robert, Narayan Naik, and Melvyn Teo. 2007. “Do
Fung, William, and David A. Hsieh. 2004. “Hedge Fund Bench- Hedge Funds Deliver Alpha? A Bayesian and Bootstrap Analy-
marks: A Risk-Based Approach.” Financial Analysts Journal, sis.” Journal of Financial Economics, vol. 84, no. 1 (April):229–264.
vol. 60, no. 5 (September/October):65–80. Kosowski, Robert, Allan G. Timmermann, Russ Wermers, and
Fung, William, David A. Hsieh, Narayan Y. Naik, and Tarun Hal White. 2006. “Can Mutual Fund ‘Stars’ Really Pick Stocks?
Ramadorai. 2008. “Hedge Funds: Performance, Risk, and New Evidence from a Bootstrap Analysis.” Journal of Finance,
Capital Formation.” Journal of Finance, vol. 63, no. 4 vol. 61, no. 6 (December):2551–2595.
(August):1777–1803. Liang, Bing. 1999. “On the Performance of Hedge Funds.” Finan-
Getmansky, Mila. 2005. “The Life Cycle of Hedge Funds: Fund cial Analysts Journal, vol. 55, no. 4 (July/August):72–85.
Flows, Size and Performance.” Unpublished paper, University Litterman, Robert B. 2004. “The Active Risk Puzzle.” Journal of
of Massachusetts at Amherst (January). Portfolio Management, vol. 30, no. 5 (September, 30th Anniver-
Gottesman, Aron, and Matthew R. Morey. 2006. “Manager Edu- sary Issue):88–93.
cation and Mutual Fund Performance.” Journal of Empirical Lynch, Anthony W., and David K. Musto. 2003. “How Investors
Finance, vol. 13, no. 2 (March):145–182. Interpret Past Fund Returns.” Journal of Finance, vol. 58, no. 5
Goyal, Amit, and Sunil Wahal. 2008. “The Selection and Termi- (October):2033–2058.
nation of Investment Management Firms by Plan Sponsors.” Massa, Massimo, and Lei Zhang. 2009. “The Effects of Organi-
Journal of Finance, vol. 63, no. 4 (August):1805–1847. zational Structure on Asset Management.” Working paper,
Grossman, Sanford J., and Joseph E. Stiglitz. 1980. “On the INSEAD (October).
Impossibility of Informationally Efficient Markets.” American Moskowitz, Tobias. 2000. “Discussion: ‘Mutual Fund Perfor-
Economic Review, vol. 70, no. 3 (June):393–408. mance: An Empirical Decomposition into Stock-Picking Tal-
Harlow, W. Van, and Keith C. Brown. 2006. “The Right Answer ent, Style, Transactions Costs, and Expenses.’” Journal of
to the Wrong Question: Identifying Superior Active Portfolio Finance, vol. 55, no. 4 (August):1695–1703.
Management.” Journal of Investment Management, vol. 4, no. 4 Pástor, Lubos, and Robert Stambaugh. 2002. “Mutual Fund
(Fourth Quarter):15–40. Performance and Seemingly Unrelated Assets.” Journal of Finan-
Hirshleifer, David, Siew Hong Teoh, and Jeff Jiewei Yu. 2011. cial Economics, vol. 63, no. 3 (March):315–349.
“Short Arbitrage, Return Asymmetry, and the Accrual Anom- Paul, Joseph Gerard. 2009. “No Time to Be Passive—Get Active
aly.” Review of Financial Studies, vol. 24, no. 7 (July):2429–2461. Now.” Research report, AllianceBernstein (January).



Phalippou, Ludovic, and Oliver Gottschalg. 2009. “The Perfor- ———. 2003. “Is Money Really ‘Smart’? New Evidence on the
mance of Private Equity Funds.” Review of Financial Studies, Relation between Mutual Fund Flows, Manager Behavior, and
vol. 22, no. 4 (April):1747–1776. Performance Persistence.” Unpublished paper, University of
Stewart, Scott D., John J. Neumann, Christopher R. Knittel, Maryland (May).
and Jeffrey Heisler. 2009. “Absence of Value: An Analysis of ———. 2010. “A Matter of Style: The Causes and Consequences
Investment Allocation Decisions by Institutional Plan Spon- of Style Drift in Institutional Portfolios.” Unpublished paper,
sors.” Financial Analysts Journal, vol. 65, no. 6 (November/ University of Maryland (May).
December):34–51. Wermers, Russ, and Tong Yao. 2010. “Active vs. Passive Invest-
Surowiecki, James. 2004. The Wisdom of Crowds. New York: ing and the Efficiency of Individual Stock Prices.” Unpublished
Random House. paper, University of Iowa and University of Maryland (May).
Wei, Kelsey, Russ Wermers, and Tong Yao. 2009. “Uncommon Winkelmann, Kurt. 2003. “Developing an Optimal Active Risk
Value: The Investment Performance of Contrarian Funds.” Budget.” In Modern Investment Management: An Equilibrium
Unpublished paper, University of Texas at Dallas, University of Approach. Edited by Robert Litterman. Hoboken, NJ: John Wiley
Iowa, and University of Maryland (November). & Sons.
Wermers, Russ. 2000. “Mutual Fund Performance: An Empirical Xiong, James, Roger G. Ibbotson, Thomas Idzorek, and Peng
Decomposition into Stock-Picking Talent, Style, Transactions Chen. 2010. “The Equal Importance of Asset Allocation and
Costs, and Expenses.” Journal of Finance, vol. 55, no. 4 Active Management.” Financial Analysts Journal, vol. 66, no. 2
(August):1655–1703. (March/April):22–30.

www.cfainstitute.org
The global membership organization that awards the CFA® and
CIPM® designations, CFA Institute leads the investment profession globally by
setting the highest standards of ethics, education, and professional excellence.

SENIOR INVESTIGATOR, INDUSTR Y MATTERS—LONDON

DIRECTOR, EXAMINATION DEVELOPMENT—HONG KONG

DIRECTOR, EXAMINATION DEVELOPMENT—LONDON

Details for these and other available opportunities can be found at
www.cfainstitute.org/careers
We offer an excellent compensation and benefits package, including medical, dental, retirement plan,
life and disability insurance, educational assistance, training and educational opportunities,
International Rotation Assignment Program, Wellness Program, Leave Program, and more.
Please respond by sending resume with cover letter
and remuneration requirements via e-mail to hr@cfainstitute.org.

EOE


Active managementmostlyefficientmarkets faj

More Related Content

What's hot (19)

Similar to Active managementmostlyefficientmarkets faj (20)

More from bfmresearch (20)

Recently uploaded (20)

Active managementmostlyefficientmarkets faj