June 06 CP.

book Page 1 Friday, June 2, 2006 12:48 PM

The Future Role of Hedge Funds

Clifford S. Asness
Managing and Founding Principal
AQR Capital Management, LLC
New York City

Hedge funds generate returns through managers’ skill (measured by alpha) as well as
managers’ systematic risk taking (measured by hedge fund beta). Hedge funds
combined with index funds are now poised to replace traditional active management.
To position themselves for such a future role, hedge funds must correct some of the
industry’s negative characteristics, including high correlations with the market, the
misuse of momentum strategies, and lags in marking to market. To appeal to institutional
investors, hedge funds must also improve their professionalism by altering the way they
make decisions, reducing their fees (or at least rationalizing them), and striving for
increased transparency.

edge funds represent the future of active man-
H agement, and that future will consist of two
complementary elements: (1) hedge funds that offer
alpha and (2) index funds that generate the bench-
mark rate of return or perhaps some alpha in a disci-
plined manner. Combined, these two approaches will
replicate and eventually replace traditional active
management. But before that future can blossom,
hedge funds must shake off certain negative charac-
teristics of their current behavior and undergo some
necessary evolutionary changes.
Investor Interest
When investing in hedge funds, investors expect a
combination of positive expected returns, alpha, and
diversification of risk through low to zero correlation
with traditional asset classes. The hedge fund world
does not always achieve such low correlations, but
certainly that is one of its primary goals. By adding
hedge funds to their portfolios, investors expect to
increase their risk-adjusted return—that is, improve
their efficient frontier—in one (or both) of two ways:
• Earn more expected return with the same or less
risk.
• Take less risk to earn the same or higher return.
This presentation comes from the Hedge Fund Management 2006
conference held in Philadelphia on 16 February 2006.
Editor’s Note: This presentation is based on two previously pub-
lished articles (Asness 2004a and Asness 2004b).
In theory, a hedge fund manager could run a
portfolio with negative expected returns by using
excellent hedging strategies that reduce risk, whereby
the diversification benefits would more than offset the
negative performance. Such a portfolio, however, is
not attractive enough to be a practical business, so I
will rule out that scenario as a likely way to use hedge
funds in the long term. Therefore, the real benefit of
adding a hedge fund as a separate asset class in a
portfolio that already includes the traditional asset
classes of equities, fixed income, and real estate, espe-
cially under current conditions, is to shift the efficient
frontier (found through optimization) to the left. In
other words, add return with the same level of risk or
reduce risk for the same expected return.
Figure 1 shows the historical P/E of the S&P 500
Index, with price divided by the rolling average of
10-year trailing real earnings. And Table 1 reveals
that periods of high P/Es precede poor inflation-
adjusted stock market performance in the following
10-year period. Given the current high levels of P/Es,
this analysis hints at the prospect of lower equity
returns in the future (and using separate data, lower
bond returns), which is one reason why investors are
turning to alternative investments. Certainly, the
prospects are borne out by the past five years of
equity returns—that is, ex post. Such a prediction was
much more dramatic at the market peak of 2000 when
only a few of us were warning of the prospects we
foresaw. Now, after several years of poor returns,

©2006, CFA Institute x cfa pubs .org JUNE 2006 x 1

June 06 CP.book Page 2 Friday, June 2, 2006 12:48 PM

CFA Institute Conference Proceedings Quarterly

Figure 1. P/E of S&P 500: Price Divided by 10-Year Real Earnings, 1880–2004
P/E
50

40

30

20

10

0
1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Note: P/Es are based on current price divided by the average of the past 10 years’ earnings adjusted
for inflation.

Table 1. P/E Range in Relation to the Real Stock Alpha is, in a sense, insider knowledge obtained
Market Return in the Next 10 Years, through legal means. Alpha represents the highly
January 1927–February 2004 idiosyncratic knowledge that hedge fund managers
Real Stock Market Return develop through their own work and insights and
in the Next 10 Years that they then use to earn returns that traditional
Median Worst management is, in theory at least, unlikely able to
P/E Range (annual) (total) earn. In the second category of strategies, hedge fund
5.2–10.1 10.9% 46.1% managers take on risks that other managers do not
10.1–11.9 10.7 32.0 want to bear or provide liquidity that other managers
11.9–14.6 10.0 4.0 want to obtain. Furthermore, for hedge funds to
14.6–17.2 7.6 –20.9 deliver pure alpha without market beta, they have to
17.2–19.9 5.3 –32.0 use leverage, derivatives, and short selling. Hedge
19.9–31.7 –0.1 –35.5 funds do not, of course, fit neatly into one or the other
of the strategy categories listed. Rather, they combine
Note: P/Es are based on current price divided by the average of the
past 10 years’ earnings adjusted for inflation. various elements of alpha, hedge fund beta, and tra-
ditional market beta as well.

Turning Skill into an Investable Diversifying

many investment professionals are making the same
argument. Thus, Table 1 presents a cautionary tale. It
does not mean that investors can find alpha only in
hedge funds or that hedge funds can, in fact, provide
alpha. It simply indicates that the prospect of lower
returns on most investments is a reason why hedge
funds today are seen as a popular alternative—even
at a time when, in all likelihood, hedge funds will also
bring lower returns than they have historically.
How Hedge Funds Make Money
Hedge funds generally fall into two broad categories
of strategies that are used to provide diversified pos-
itive expected return:
1. investment skill strategies, or alpha, that are
turned into investable diversifying assets and
2. nontraditional arbitrage strategies, or hedge fund
betas, that provide liquidity or take a risk for
which they are paid a premium.
Asset. Traditional active management (as opposed
to hedge fund management) can be defined as
Active management = Index
+ (Active management − Index).
Similarly, a simple hedge fund can be defined as
Hedge fund = Cash + (Active management − Index).
The result is a pure alpha hedge fund. If the beta of
the active portfolio to the index is 1.0, the hedge fund
is uncorrelated with the index. If the alpha is positive,
the fund earns a positive expected return. If the alpha
is idiosyncratic to the manager, investors cannot get
that alpha anywhere else. The result, however, is
something akin to a skillful manager who is picking
good stocks but taking, say, only a 2 percent tracking
error from the index. I have thus created a hedge
fund—but a very boring hedge fund.
Add leverage, however, and the situation
changes dramatically—and becomes a lot less boring.

2 x JUNE 2006 ©2006, CFA Institute x cfa pubs .org

June 06 CP.book Page 3 Friday, June 2, 2006 12:48 PM
A hedge fund manager can lever (or de-lever) the
difference, as shown in the following:
Hedge fund = Cash + Leverage
× (Active management − Index).
Levering makes the fund riskier, but it does not nec-
essarily raise the degree of risk unacceptably. If I
assume that skill exists and is identifiable and that
leverage is greater than 1, the hedge fund should
charge fees larger than those charged by the corre-
sponding active manager. For example, if leverage
equals 2, the fair fee for the hedge fund is double the
fair fee for the same amount of dollars invested with
a traditional active manager because the hedge fund
manager is providing double the active manager’s
skill. One could argue that risk is bad and that the
hedge fund manager is taking double the risk of the
active manager. But under the assumption that the
hedge fund manager is offering double the positive
return that the active manager is offering, then dou-
bling the fee is fair. Think of the situation in terms of
a finite capacity of alpha such that fees should be
measured per unit of expected alpha.
Note that the construct “Active management –
Index” imposes an arbitrary constraint on shorting,
which, if lifted, can further improve returns, under
the assumption that skill is present.
Using Arbitrage Strategies or Hedge Fund
Betas. Many years have passed since I was a PhD
student, but the academic in me still recoils at the
misuse in our industry of the word “arbitrage.” When
I left the University of Chicago 15 years ago, arbitrage
meant “riskless profit.” This definition does not work
in the world of actual practice, but certainly, it has the
virtue of clarity. In common industry usage, unfortu-
nately, the term has come to mean something akin to
a bet that is liked, which I find almost entirely unac-
ceptable. Still, in actual practice, an arbitrage strategy
does some or all of the following:
• It goes long and short similar securities while
hedging out unwanted risks, such as general
market risk.
• It is based on the assumption that mispricings or
risk premiums are economically significant and
that the fund will produce excess return, includ-
ing transaction costs.
• It frequently generates excess returns from posi-
tive carry positions in a normal environment
(such arbitrage trades are not riskless).
• It includes an insurance characteristic in that it
assumes a large risk or many small risks that are
unwanted by others (such risks are sometimes
nonlinear).
• It is affected by “flows,” which may cause the
arbitrage strategy to become a common risk fac-
tor even if it is usually idiosyncratic.
©2006, CFA Institute x cfa pubs .org
The Future Role of Hedge Funds
Note that these strategies are different from skill
because they are activities that, once described, can
be emulated by any investment manager willing to
use them. In contrast, I would argue that when an
arbitrage strategy, such as merger arbitrage, was first
introduced, it was a true form of alpha. Thus, alpha
can become a hedge fund beta as it becomes better
known and used by others. Alpha of an increasingly
popular strategy does not necessarily fall to zero, but
the related strategy will generate a much lower
Sharpe ratio with more systematic risk.
Exhibit 1 presents several kinds of hedge fund
strategies and describes their essential characteris-
tics along with the hedge fund beta and alpha that
each offers.
Hedge Funds as an Evolutionary
Step
When operating properly, hedge funds provide
investors with management skill and risk or liquidity
premiums as separate investable assets. By contrast,
traditional active management is a package deal
largely dominated by stock pickers that typically hug
the index, which, of course, negates the reason why
investors seek out active management. My conten-
tion, therefore, is that the investment world of the
future will look like a combination of hedge funds
and index funds; hedge funds will provide the
advantages that traditional active management is
supposed to provide, and index funds will provide
what many stock pickers have, in fact, reverted to
offering but at a lower fee. Such a restructuring of the
investment world offers considerable advantages
over traditional active management.
First, many hedge fund strategies cannot be
implemented in a traditional structure. For example,
managers cannot include merger arbitrage in a long-
only portfolio. Therefore, hedge funds offer a source
of return that investors want in their portfolios but
cannot have within the traditional structure. In addi-
tion, certain hedge fund strategies make most inves-
tors in the market better off as a whole because most
investors benefit when hedge funds are properly
used to help provide liquidity or bear risks other
investors wish to avoid.
Second, hedge funds offer fee clarity with cleaner
performance attribution. Traditional active manage-
ment is a tie-in sale, meaning that investors have to
figure out their active managers’ betas and tracking
errors and determine for themselves how the fees are
being divided between each source of return. Fee
clarity and performance attribution can be cleaner in
a hedge fund.
JUNE 2006 x 3
June 06 CP.book Page 4 Friday, June 2, 2006 12:48 PM

CFA Institute Conference Proceedings Quarterly

Exhibit 1. Example Hedge Fund Strategies with Their Betas and Alphas
Potential Systematic Profit Source
Hedge Fund Strategy Hedge Fund Activity (hedge fund beta) Potential Manager Alpha
Convertible arbitrage Long embedded option from The market systematically pays the Better models of rich/cheap and
convertibles hedged with stock. fund for taking on unfamiliar or hedge ratios.
uncomfortable convertible.

Merger arbitrage Long a target and short an acquirer The market systematically pays the More accurate underwriting
where spread is not fully closed. fund to provide insurance that deals through better deal selection,
close (i.e., the average spread better risk management, and
overcomes the failures). systematic deal selection.

Statistical arbitrage Long and short a hedged stock The market systematically pays the Low-cost trading, better risk
portfolio based on short-term fund for providing short-term systems for removing
supply and demand anomalies. insurance and liquidity, perhaps unintended bets, and other
against large information trades. short-term factors.

Equity market neutral Long–short under-/overpriced Expected returns on cheap stocks Momentum and other factors to
stocks over intermediate term. exceed those on expensive stocks. attempt to improve timing
Tends to be more quantitative. aspect of value strategy.

Long–short equity Long–short stocks based on valua- Does not have a nontraditional Manager expertise, acumen,
tion, catalysts, forensic accounting, systematic profit source. Often and effort.
and so on. Tends to be more funda- includes stock market beta.
mental than quantitative.

Finally, traditional active management can be
viewed as a strangely constrained combination of a
particular type of hedge fund and index fund invest-
ing. Traditional management jams together two things
and does not allow them to be separate, whereas
everything that can be done in a traditional structure
can be done through a hedge fund plus an index fund
but without the constraint. And unconstrained opti-
mization beats constrained optimization because con-
straints ultimately shift the efficient frontier to the
right, thereby reducing rather than improving the
trade-off between expected return and expected risk.
Until about a year ago, major newspapers
defined hedge funds as lightly regulated investment
vehicles, catering to the wealthy, that make relatively
short-term bets on the direction of stocks. If hedge
funds hope to continue moving beyond the realm of
“catering to the wealthy,” they must make a change
that will inspire confidence beyond this niche mar-
ket. They must not only accept but also embrace a
certain degree of institutionalization. Many hedge
fund managers cringe at such a suggestion because
they define institutionalization as a creativity-
draining bureaucracy that saps hedge funds and
hedge fund managers of all they were meant to be.
And certainly, there is some truth in such a defini-
tion. Too many strictures can remove the idiosyn-
cratic nature that characterizes hedge funds and that
allows them to do their best work. But I would argue
for a more benign definition in which institutional-
ization represents a needed professionalism that
must happen if hedge funds are going to fit into
modern portfolios with enough capacity to matter.
The truth is that institutional portfolios will not fully
embrace hedge funds until they can perceive a
degree of professionalism that they find acceptable.
Problem Areas. To achieve a status of greater
significance in the investment world, hedge funds
must address a number of serious issues, such as
• hedge fund betas/correlations,
• momentum strategies,
• lags in marking to market, and
• hot money.
■ Hedge fund betas/correlations. For hedge funds
to provide skill and risk as separate investable assets
with clear performance attribution, hedge funds must
run at or near zero beta; they cannot be closely corre-
lated with the market. If hedge funds do not short a
lot and are correlated with the market, they are no
different from traditional mutual funds, at least in the
attribution of performance. The most egregious
offenders in this case are long–short equity hedge
funds, although other subcategories offer fairly dra-
matic evidence of market correlation. Figure 2, Panel
A, compares the rolling one-year returns from
December 1994 through December 2004 of the S&P
500 with long–short equity managers, and Panel B
compares the monthly results in 2004 for the S&P 500
with long–short equity managers.
The strong positive correlation is obvious. For
instance, Panel B shows that in 2004, the S&P 500
began well, dipped throughout much of the middle
of the year, and then had a strong positive response
after the U.S. presidential election, which saved the
stock market. The story for long–short equity funds
is precisely the same. Investors are taking on far more
risk than they realize if they believe that such hedge

4 x JUNE 2006 ©2006, CFA Institute x cfa pubs .org

June 06 CP.book Page 5 Friday, June 2, 2006 12:48 PM

The Future Role of Hedge Funds

Figure 2. CSFB/Tremont Long–Short Equity Funds vs. S&P 500

A. Rolling One-Year Returns over Cash,
December 1994–December 2004
Return (%)
50
40
30 S&P 500
20
10
0
−10 CSFB/Tremont
Long/Short Equity Funds
−20
−30
−40
12/94 12/95 12/96 12/97 12/98 12/99 12/00 12/01 12/02 12/03 12/04

B. Cumulative Return over Cash,

December 2003–December 2004
Return (%)
12

10

CSFB/Tremont
4 Long/Short Equity Funds

2
S&P 500
0

−2
12/03 1/04 2/04 3/04 4/04 5/04 6/04 7/04 8/04 9/04 10/04 11/04 12/04

funds are either uncorrelated or have low correla-
tions with traditional asset classes.
■ Momentum strategies. This same reasoning
can be extended to hedge fund momentum strategies.
Start with a standard regression equation:
L – S equityt = α + β × S&P 500t ,
where L–S is long–short. Then, run the single regres-
sion model over the monthly sample period from
January 1994 to August 2004:1
L – S equity = 0.36% + 0.48 × S&P 500, R ² = 35.7%.
(1.48) (8.19)
This is the old market model: Regress long–short
equity funds monthly and use about 10 years of data
on the S&P 500 monthly return. The beta coefficient of
0.48 is statistically significant given its high t-statistic,
thus indicating either underinvesting or shorting by
the hedge fund but not to a beta of zero. The conclu-
sion is that these hedge funds have beta, but it is not
dominating what they do; they are clearly doing
something else as well. The alpha is positive but not
statistically significant.
Next, develop a simple model for beta. Instead of
beta being a number, it is now itself an equation:

1Returns are quarterly overlapping excess returns over T-bills. The

β = γ + δ × TRENDt ,
t-statistics reported in parentheses are adjusted for overlapping where TRENDt is the prior one-year performance of
observations. the S&P 500 ending before quarter t. The hypothesis

©2006, CFA Institute x cfa pubs .org JUNE 2006 x 5

Asness.X.fm Page 6 Friday, June 2, 2006 1:24 PM
CFA Institute Conference Proceedings Quarterly
is that delta (the constant term in the regression) is
positive and beta (the coefficient of TRENDt) is higher
when the S&P 500 is higher and that hedge fund
managers follow momentum strategies in their betas.
Now, instead of regressing the long–short equity
managers solely on the S&P 500, the following equa-
tion also regresses on the product of the S&P 500 and
the prior one-year S&P 500 return, which gives an
estimate of delta, as shown:
L – S equityt = α + γ × S&P 500t + δ × TRENDt
× S&P 500t
L – S equity = 0.21% + 0.46 × S&P 500t + 1.18
(0.93) (12.86) (4.69)
× TRENDt × S&P 500t , R² = 42.0%.
The adjusted R2 shows a modest improvement in
the explanatory power of this multiple regression. The
coefficient of TREND is positive with a statistically
significant t-statistic of 4.69. Thus, the probability that
hedge funds are not following TREND strategy in
setting their betas is low. Based on this analysis, long–
short equity managers appear, after strong years for
the S&P 500, to be fully invested. To illustrate this
finding, a fitted beta using this equation varies
between about 0.1 and 1.0, not even straddling the
value of zero. The intriguing result is that long–short
managers, in fact, have become closer to long-only
managers, particularly after bull markets.
Momentum strategies in beta are not necessarily
bad for the investor. They are a form of option repli-
cation and portfolio insurance. They work well in
Figure 3. CSFB/Tremont Distressed Funds vs. S&P 500: Cumulative Return
over Cash, December 2003–December 2004
Return (%)
14
12
10
8
CSFB/Tremont
6 Distressed Funds
2
S&P 500
0
−2
12/03 1/04 2/04 3/04 4/04
6 x JUNE 2006
long, slow bear markets (such as in 2001 and 2002)
but poorly in short, sharp bear markets (such as the
stock market crash of 1987) when returns are oscillat-
ing rather than trending in one direction. Momentum
strategies have, in fact, been a small positive for the
past 11 years. I use momentum strategies in some of
my own modeling, so I am not opposed to them.
Hedge funds that are net momentum rather than
contrarian can make money. But for the market as a
whole, they can intensify financial crises, which is
part of the dark side that results when hedge fund
strategies are misapplied.
■ Lags in marking to market. To illustrate the lags
in marking to market, I have chosen the distressed
index shown in Figure 3. The graph shows the cumu-
lative return over cash for distressed funds and the
S&P 500 for each month in 2004. Note that the upper
line, representing the cumulative performance of dis-
tressed funds, does not look like any security on
earth; securities go up and down, but the distressed
funds do nothing but go up. The steadily increasing
line related to the performance of the distressed index
seems like something that perhaps has already gone
up in value and is being slowly and surely marked
up in value as time goes by, which may not be bad
unless the investor is running a regression with
monthly numbers on the hedge fund manager’s
returns and decides the manager is riskless. Investors
who do not think such practices occur are naive. This
type of practice occurs quite frequently, and the out-
come is a portfolio that superficially looks far less
risky than it really is.
5/04 6/04 7/04 8/04 9/04 10/04 11/04 12/04
©2006, CFA Institute x cfa pubs .org
June 06 CP.book Page 7 Friday, June 2, 2006 12:48 PM
This behavior is economically significant, and I
co-authored a paper titled “Do Hedge Funds Hedge?”
(Asness, Krail, and Liew, 2001) to offer a perspective
on the problem. Over-the-counter strategies demon-
strate the worst instances of the problem, and tactical
futures trading shows the fewest instances, but that
does not mean that tactical futures managers are more
virtuous than other hedge fund managers. I am sure
they would love to have lags in marking to market.
But tactical futures strategies do not offer opportuni-
ties for lags. My advice to anyone investing with a
hedge fund is to run regression tests on past condi-
tions because the result of lags in marking to market
is that almost all aggregate hedge fund risk estimates
are understated, both for volatility and beta.
■ Hot money. Hot money leads to a variety of
problems. Hedge funds do their best work when they
take risks, acting as investors of last resort when things
are cheap and getting fair price for buying invest-
ments that everyone else is afraid of. Such behavior is
easier to maintain in a world of stable investors than
in one of hot money. Hot money induces transaction
costs for longer-term investors. It also induces manag-
ers to think too much about the short term and take
too little risk. In a rational market where hedge funds
are paid to provide liquidity and take risks, they
should have the steadiest hand, but hot money tends
to weaken their hand and make them less effective at
what they are supposed to do best.
Future Evolution. Besides dealing with the seri-
ous issues mentioned earlier, hedge funds must
evolve if they expect to be significant players in the
future investment world. They must be prepared to
• moderate investors’ expectations of future per-
formance,
• address questions about capacity,
• establish rational fees,
• educate clients to tolerate (and understand)
headline risk, and
• establish reasonable transparency guidelines.
■ Expectations and capacity issues. The first two
evolutionary steps are linked so closely that they
must be dealt with together. Regarding expectations,
investors (and managers) must acknowledge that
one of the few things known for certain about capital
markets is that as more money goes into a strategy or
asset class, future returns drop. Hedge fund manag-
ers and investors need to prepare for this reality. As
for capacity, the question should never be, “What
capacity exists?” Rather, it should be, “What capacity
exists at a certain level of risk-adjusted return?” In
response to questions like these, institutions are per-
forming calculations, and they are starting to realize
that they do not need a Sharpe ratio of 1 or even as
high as 2. Instead, a prospective asset class with a
Sharpe ratio of 0.75, 0.50, or even 0.25—if truly uncor-
©2006, CFA Institute x cfa pubs .org
The Future Role of Hedge Funds
related with the portfolio—can substantially
improve the risk-adjusted returns of a portfolio of
traditional assets if the prospective asset class offers
diversification benefits. Institutions, therefore, will
invest until they get those lower Sharpe ratios
because institutions set the price and expected return.
■ Rational fees. Although the current hedge
fund market is still a seller’s market, fees are bound
to come down. Institutions and their growing pres-
ence as hedge fund investors may not drive down
hedge fund fees precipitously, but they should force
them to more rational levels. For example, institutions
will still be willing to pay well for alpha-like idiosyn-
cratic skill but they will pay less for systematic expo-
sure to known nontraditional strategies—although
they will pay more than they would for traditional
stock market beta (index fund fees) because hedge
fund skills are still rarer than index fund skills. I
foresee an increased role for fixed fees for what insti-
tutions consider hedge fund beta, as well as an
increased use of innovative structures—such as
longer lockups instead of higher fees to ration scarce
capacity with performance fees paid at the end of the
lockup. Finally, I believe there will be a careful exam-
ination of the pros and cons of performance fees.
■ Headline risk. Headline risk is an institution’s
worry that an embarrassing public incident—such as
a major failure of a hedge fund’s strategy or a charge
of fraud or malfeasance by hedge fund managers—
could occur and affect the institution’s reputation.
Although headline risk almost certainly gets more
attention than it deserves, it is still one of the elements
that slows the flow of money into hedge funds. There-
fore, an important goal for all hedge fund managers
should be to remind their institutional clients of the
reasons why they invest in hedge funds: Relaxed
constraints allow hedge fund managers to do more
things, take more risk, and use more tools to achieve
their special purposes. They need to also remind
clients that headline risk applies to extreme situations
and that they need to judge hedge funds by their effect
on a portfolio, preferably over time, and not by a
moment’s worth of fleeting negative publicity.
■ Transparency guidelines. Transparency remains
a burning issue for hedge funds, one that is especially
important to institutional clients. Hedge funds were
originally developed in an atmosphere of secrecy
regarding their proprietary knowledge, and this
atmosphere still affects hedge fund managers’ think-
ing. In particular, hedge fund managers fear revealing
their proprietary strategies or their proprietary posi-
tions, especially when those positions are vulnerable.
The classic example of the second circumstance occurs
when the hedge fund manager has an illiquid short
and could get squeezed if others knew of the position.
JUNE 2006 x 7
June 06 CP.book Page 8 Friday, June 2, 2006 12:48 PM
CFA Institute Conference Proceedings Quarterly
For the most part, however, I believe these con-
cerns are overstated and that the real reason why
hedge fund managers today avoid transparency is
simple logistics. Providing transparency would
require hedge fund managers to make significant
administrative changes in their normal mode of oper-
ation. It would mean letting clients into the back
office, perhaps several times a day, and opening man-
agement to cross-examination over any and all trades
and positions. It would also mean making portfolio
managers available for such questioning when the
fund would rather have them doing the work they
were hired to do. Furthermore, I believe hedge fund
managers see additional transparency as pointless
because they believe the information provided
would simply go into an institutional black hole
where it would lie unused. And finally, I believe that
in their hearts, hedge fund managers fear that trans-
parency would reveal the basic simplicity and lack of
magic in what they do.
Ultimately, I think reasonable compromise posi-
tions are available. I believe that hedge fund manag-
ers can easily provide summary risk reports that
reveal useful aggregates rather than individual posi-
tions. I also believe that some process transparency
can be offered without revealing proprietary infor-
mation. If hedge funds want to expand their presence
in the market, they will find themselves dealing more
regularly with institutions, and institutions expect
REFERENCES
Asness, Clifford S. 2004a. “An Alternative Future: Part II.” Journal
of Portfolio Management, vol. 31, no. 1 (Fall):8–23.
———. 2004b. “An Alternative Future.” Journal of Portfolio Manage-
ment (30th Anniversary Issue, September):94–103.
8 x JUNE 2006
transparency. Consequently, hedge funds must learn
to adapt to the clients of their future.
Conclusions
Following are the main ideas to take away from this
presentation:
• Hedge funds should exist to take sources of
expected return, such as idiosyncratic skill (alpha)
and nontraditional arbitrage (hedge fund beta),
and turn them into separate investable assets.
• Hedge funds should not be an expensive stock
market index fund with a few shorts thrown in
for credibility.
• Not only are hedge funds here to stay; the market
model of the future could likely be a combination
of index funds and hedge funds.
• Institutionalization of hedge funds, with its pros
and cons, is definitely under way, although best
practices for both investors and managers
require serious improvement.
• Risks and issues abound in traditional markets
(long-only stocks and bonds), so if a time ever
existed to look for nontraditional sources of
return, the time is now.
• Hedge funds are the future of active manage-
ment, but many things have to change before this
future arrives.
This article qualifies for 0.5 PD credits.
Asness, Clifford S., Robert Krail, and John Liew. 2001. “Do Hedge
Funds Hedge?” Journal of Portfolio Management, vol. 28, no. 1
(Fall):6–19.
©2006, CFA Institute x cfa pubs .org
June 06 CP.book Page 9 Friday, June 2, 2006 12:48 PM
Question and Answer Session
Clifford S. Asness
Question: Absent some exoge-
nous event, what normal market
conditions might create a shift in
the balance of power back to inves-
tors from hedge fund managers?
Asness: Just being boring, in
terms of performance, for a long
time could have a tremendously
negative effect on hedge funds.
And I define boring as being a
return anywhere between the max-
imum of the S&P 500 return and
zero. Of course, if the S&P 500 were
to be negative, then hedge funds
could be boringly positive and still
be okay.
Question: If hedge funds do, in
fact, deliver a disproportionate
amount of alpha relative to tradi-
tional long-only funds, could we
actually see fees increase over time?
Asness: No. I think hedge fund
fees are past the point of being sup-
ported by the argument that hedge
fund alpha is that much better.
Question: If alpha can become
beta over time and if an investor
believes the markets have chronic
inefficiencies, can you make the
argument that an investor focused
on chronic inefficiencies cannot
reliably separate alpha from beta?
Asness: Observers are seldom
likely to look at market results and
agree that this is alpha or that is
beta, so your starting point is not as
obvious as you might think. Each
observer is likely to come up with
a different answer.
The most important insight to
take away from this part of the dis-
cussion is the concept of frame-
work. Your answer does not have
to be exactly like mine, but if you
have a framework for judging mar-
ket phenomena, then you are more
likely to come up with a reasonable
answer. Furthermore, your answer
will change through time, which
means you cannot just conduct a
©2006, CFA Institute x cfa pubs .org
10-year backtest and say this is
alpha and that is beta because what
was alpha 10 years ago is beta
today. Still, just being forced to ask
the questions, “Is this alpha or beta,
and is my answer today different
from what it was last year?” is a
useful exercise.
Question: What is your sense of
the expected standard deviation
of returns for market-neutral
strategies?
Asness: Market neutral can
mean almost anything, but most
hedge fund managers mean single-
digit rather than double-digit
volatility—say, 3–7 percent volatil-
ity. But you could find someone
who would see market neutral as
being quite different, so do not
assume that everyone is coming
from the same starting point.
Question: How does an inves-
tor assess what the expected
alpha would be for any particular
strategy?
Asness: The framework helps
you ask the right questions. At the
end of the day, you still have to sit
down with the hedge fund manag-
ers and determine if they are hedg-
ing out their market risk, and if
they are, are they people who
understand something that the
market does not? Are they digging
up information that the market
does not have? Do they have a
mental or a literal model for the
value of a company that is superior
to the one being used in the mar-
ket? Are they getting information
more quickly (in a legal manner, of
course) than the market gets it?
Asking the right questions and
determining the actual answers is
the hardest part.
Furthermore, we have a prob-
lem with garbage in/garbage out
because two sets of skills are
needed for the successful use of
hedge fund strategies. First, the
Q&A: Asness
managers must have skill, and sec-
ond, the investor must have skill at
picking the managers. If the inves-
tor cannot pick good managers, he
or she should go to a passive port-
folio. We all know how to do it in
the long-only world. No one gets
fooled by beta in the long-only
world. You never pick a manager
who has lost to the index from 1995
through 1999 but brags to you
about how much he or she is up.
Question: Because hedge fund
returns are not normally distrib-
uted, does it make sense to use
Sharpe ratios? Assuming not, are
there any other measures you
believe to be helpful?
Asness: I am a big fan of simple
measures. There are some well
thought out downside risk mea-
sures, such as Sortino measures. A
lot of the issue with nonnormality is
that it is a fancy way of saying some
scary left tail is out there. The distri-
bution of historical returns looks
normal most of the time, but occa-
sionally, everything falls to pieces.
It is much more productive to
stick with some of the basic mea-
sures. Use the back of the envelope;
use your acumen to determine if
you think there is left-tail risk to a
strategy, and then estimate what
you think that will cost you. It is a
“guesstimate,” but we have noth-
ing better than guessing. Think
about what the risk will cost you,
on average, every 20 years, and
then guess how much that would
reduce your real Sharpe ratio if it
were just a dollar loss.
Working with complicated
measures scares me because I think
they are too much of a crutch. They
are dependent on empirical data,
and they are heterogeneous. Some
very smart people use complicated
measures, but I believe in the sim-
ple ones and then thinking about
what the left tail looks like.
JUNE 2006 x 9