Understanding Expected Returns

Antti Ilmanen
Managing Director
AQR Capital Management (Europe) LLP
London
Investors tend to think of expected returns as a function of asset class risk, but this thinking
may have led them to take on too much equity risk. For behavioral reasons, diversifying across
investment styles, such as blending momentum and value, may offer greater returns for less
risk. Limited market timing may also increase returns.
❚■❚

will talk about expected returns, but I should they are, typically 90 percent or more of their port-
I make it clear that I am talking about very long-
term expected returns—nothing about current mar-
folio risk comes from equities, which is something
that many investors want to reduce because the
kets, the eurozone crisis, or the impending reces- roller coaster has been quite violent in the past 10
sion. Compared with most commentators, having years. But it is difficult to reduce the risk, especially
neutral views for the near term actually makes me when everybody wants to do it at the same time.
appear to be a raving optimist among the current Investors are in a decade of low expected returns
bearish consensus. and heightened risk, so it is natural that many want
This presentation is based on my book, Expected to reduce risk.
Returns: An Investor's Guide to Harvesting Market There are three classic ways of reducing risk.
Rewards (Ilmanen 2011). I decided to cover this huge One is a move toward riskless assets. Another one
topic—the book is 500 pages long—because I have is insurance. The third one is diversification. Unfor-
dealt with so many different asset classes over the tunately, investors do not know anymore whether
years. My goal in writing the book is to help improve a riskless asset exists. If they do find a riskless asset,
the marketplace and the investor experience. then they certainly will not earn much; and in fact,
they are earning a negative real expected return.
Big Picture for Investors Insurance is very expensive now because of its pop-
In the Indian fable of six blind men and an elephant, ularity; consider the growing number of tail risk
each man touches only part of the animal and each products being offered. Finally, the power of reduc-
has a different idea of what the object is—for exam- ing risk with diversification is being challenged.
ple, a wall of mud, a spear, a rope, a serpent, a fan, Correlations are heightened in the current environ-
or a tree trunk. Each is partly right, but all are very ment, but I would say diversification is still the best
wrong. The moral of the story is that people should choice. My answer to the question of what investors
look at multiple perspectives when thinking about should be doing now is that they should diversify
anything and, in this case, when thinking about aggressively to get away from portfolios that have
expected returns. Investors need to look beyond concentrated equity market directional risk.
historical average returns; they should also look at The main theme of my book is the importance
theories and forward-looking indicators. of harvesting multiple premiums to form a more
When applying this idea to the current difficult balanced portfolio. To find this balance, a good
market situation, it is clear that many investors’ starting point is to identify the return sources that
portfolios are dependent on equity market direc- have worked well over long time periods. That pro-
tion. No matter how well diversified investors think cess includes not only analyzing various types of
asset class premiums but also looking beyond them.
Editor’s Note: The views and opinions expressed herein are those of the I will talk about the returns for different asset classes
author and do not necessarily reflect the views of AQR Capital Man- and investing styles as well as time-varying returns.
agement, LLC, its affiliates, or its employees. Diversification does not
eliminate the risk of experiencing investment losses.
The five styles that I emphasize are value, carry,
trend or momentum, volatility, and liquidity. All of
This presentation comes from the Fourth Annual European Investment these return sources have some time variation in
Conference held in Paris on 2–3 November 2011 in partnership with expected returns, and investors should try to take
CFA France. advantage of those differences. They should not

©2012 CFA Institute  cfa pubs .org JUNE 2012  55

CFA Institute Conference Proceedings Quarterly

look just at the long-run historical average and think Panel B shows various Treasury maturities. The
that it will always be the same. long history shows that investors were rewarded by
I will recommend a certain amount of contrar- duration extensions of up to two to five years but
ian timing, but it should be done modestly. Timing then received nothing extra for further extensions.
is a concentrated bet; it is a risky activity and should It is, of course, a different story during the last 20
not be the bulk of what investors do. years, when investors received windfall gains from
falling yields. Panel C turns to the reward of bearing
credit risk. Considering returns as a function of
Finding Returns from Multiple credit ratings, average returns up to a rating of BB
Asset and Style Sources show improvements but then the premium
Table 1 shows the historical performance of many declines. The most speculative investments often
key return sources, both asset class premiums and give surprisingly poor rewards—a theme that con-
tinues to come up.
style premiums, over a 20-year history and a longer
available history. All numbers are excess returns, Value. Value stocks (those with low price-to-
thus excluding cash income. No trading costs are book multiples) have outperformed both the market
subtracted, which helps certain high-turnover strat- and growth stocks (those with high price-to-book
egies. The last column shows when the time series multiples) over many decades in all markets stud-
started. Note that the premium of U.S. equities over ied. Value also works when selecting countries and
U.S. bills was 5.2 percent between 1900 and 2009, but sectors, and it works in other asset classes. There
it is probably lower now. may be risk-based explanations for this relationship;
For bond markets, the premiums for both lon- the worst time for value is during a deflationary
ger terms and different credits are much smaller recession, such as in the 1930s or in 2008.
than the premiums in the equity markets. To get a Various behavioral interpretations, however,
return similar to equities, investors have to use may offer more compelling explanations for the out-
leverage in these markets. performance of value stocks. Here is the main narra-
The lower part of Table 1 shows the various tive. Value investing works within markets and in
types of dynamic strategy premiums—value, carry, many other contexts because of the overpricing of
trend or momentum, volatility, and liquidity—gen- the hope for growth. If there is high growth in a stock
erated over the long run, and they compare favor- or a sector or a country, investors tend to extrapolate
ably with the equity premium. The history for the further subsequent growth, resulting in high valua-
styles is not quite as long as the history for the asset tions. Typically, this expectation is followed by dis-
classes, but the table suggests that investors can appointment in the growth rate and the return.
balance their equity risk premium concentration Here is one macro example. From 1988 to 2009,
with some of these strategy alternatives. the region that showed the strongest equity returns
There are caveats, of course. One is that these was Latin America, which produced returns of 18.8
results may not be sustainable. Another important percent, despite just 2.8 percent real GDP growth.
point is that these strategy premiums typically have During this period, Asia (excluding Japan) experi-
lower volatility than equity markets. If an investor enced much faster real growth but did not offer the
combines them with equities without any shorting highest equity returns. The reason Latin America
or leverage, he or she is still going to have a very did well was because valuations were very cheap in
the late 1980s. That region had just emerged from its
high equity market direction in the portfolio. Inves-
lost decade, and starting valuations matter.
tors have to accept having some amount of leverage
to really balance their portfolios properly. Many Carry. Carry investing involves selling low-
institutions cannot do that, so they remain with yielding assets to buy high-yielding assets.
concentrated equity risk. Although this activity is best known in currency
Figure 1 shows the premiums of various asset markets, strategies that seek carry work in almost
classes, Treasury maturities, and credit ratings. The every asset class and context studied.
line shows the long-range realized premiums, and From 1993 to 2000, carry-seeking strategies gen-
the 20-year realized premiums are shown in the erated excess returns in both fixed-income and cur-
bars. Panel A shows the typical story, which is that rency market strategies. The strategy generates
risk taking is rewarded when investors go from bills stronger performance when executed across coun-
to bonds to corporates to equities to small-cap stocks tries rather than within one market. Although carry
to value stocks. These asset classes offer higher has worked almost anywhere, it does sometimes suf-
returns over long histories. fer. The rare but large losses tend to be concentrated

56  JUNE 2012 ©2012 CFA Institute  cfa pubs .org

©2012 CFA Institute  cfa pubs .org

Table 1. Historical Long-Term Returns from Static Risk and Active Strategy Premiums
1990–2009 Long History (ending in December 2009)
Compound Avg. Sharpe Compound Avg. Annualized Sharpe
Return Ratio Return Volatility Ratio Start Year
Static risk premiums
World equity premium vs. U.S. T-bill 2.0 0.23 4.5 17.3 0.35 1900
U.S. equity premium vs. U.S. T-bill 4.4 0.32 5.2 20.1 0.37 1900
World term premium vs. U.S. T-bill 3.6 0.54 0.7 8.4 0.11 1900
U.S. term premium (7–10 year vs. T-bill) 3.4 0.57 1.4 6.7 0.26 1952
U.S. IG corporate credit premium vs. Treasury 0.2 0.07 0.3 4.4 0.08 1926

Dynamic strategy premiums

Value (global equity selection) 3.6 0.52 4.6 7.2 0.68 1975
Carry (currency G–10 selection) 6.3 0.76 6.7 10.3 0.67 1978
Trend (commodity trend following) 9.2 0.88 10.2 12.1 0.85 1961
Volatility selling (equity index) 4.2 0.35 4.2 15.3 0.35 1990
Carry-seeking composite 13.0 1.07 6.9 9.8 0.76 1970
Trend-following composite 8.4 1.05 9.6 10.2 0.99 1961
Bet against beta compositea 4.4 1.18 8.5 8.9 0.95 1965
Liquidity risk factor in stocksb 6.7 0.54 5.1 12.3 0.47 1968
aFrazzini and Pedersen (2010).
b
Pastor and Stambaugh (2003).

Understanding Expected Returns

Note: All returns are computed without subtracting trading costs, but all are excess returns (i.e., not including cash return).
Sources: Remaining numbers based on data from Bloomberg; Bank of America Merrill Lynch; Barclays Capital; CRSP; Citigroup; Dimson, Marsh, and Staunton (2010); Ibbotson Associates
(Morningstar); Ilmanen (2011); and Kenneth French’s website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/).
JUNE 2012  57
CFA Institute Conference Proceedings Quarterly

Figure 1. Realized Returns of Various Asset Classes, Treasury Maturities,

and Credit Ratings for 20 Years and Long Term

A. Asset Classes
Return (%, geometric mean)
16
14
12
10
8
6
4
2
0
Short Long Long Equity Small Small
Treasury Treasury Corp. Market Caps Value
Bill Bond Bond

B. Treasury Maturities
Return (%, geometric mean)
9
8
7
6
5
4
3
2
1
0
0–3 6–9 1–3 3–5 5–7 7–10 10+
Month Month Year Year Year Year Year

C. Credit Ratings
Return (%, geometric mean)
12

10

0
Treasuries AAA/AA A BBB BB B CCCa

1990–2009
A. 1926–2009
B. 1952–2009
C. 1973–2009

a
Uses B returns before 1985.
Sources: Based on data from Bloomberg, Bank of America Merrill Lynch, Barclays Capital, CRSP,
Citigroup, Ibbotson Associates (Morningstar), and Kenneth French’s website. In the top chart, the three
U.S. fixed-income indices are from Ibbotson Associates (Morningstar), while the value-weighted U.S.
equity market index and its two subsets are from Kenneth French's website.

58  JUNE 2012 ©2012 CFA Institute  cfa pubs .org

 Understanding Expected Returns

during bad times. With carry, it seems that alpha has recent winners, but the difference is a matter of time
morphed into beta over time; it is an attractive strat- horizon. It seems that winners tend to persist in
egy, but it is also risky. performance for up to a year, and after that, a rever-
sal effect takes over. The bad news for many inves-
T r e n d a n d M o m e n t u m . Trend following
tors is that their own behavior tends to be wrong.
refers to market timing with one asset at a time.
They often chase returns on multiyear horizons,
Momentum strategies typically refer to long–short
which is when the reversal effects dominate. Rever-
strategies, such as buying stocks that were last
sal effects happen with asset classes and with man-
year’s winners and selling stocks that were last
ager selection. Moving synchronously with the
year’s losers. A distinction does exist between time-
series strategies, such as trend following, and classic medium-term crowd is one of the key behavioral
cross-sectional strategies, such as long–short mistakes that investors make.
momentum trading, because only the former take Figure 2 shows the returns for a composite
directional net exposure. carry strategy and composite trend-following strat-
A simple strategy of buying an asset that has egy in the worst 15 months for global stocks during
been going up in the last year or selling one that has 1985–2009, which also happens to be a 5 percent tail
been going down has added value in many contexts. of worst returns. The results show that trend-
This strategy has given good long-run results in following strategies made money (4 percent, on
commodity futures, equity country indices, interest average) in 13 of the 15 months and the carry strat-
rate futures, and currencies. It may seem contradic- egies lost (–5 percent, on average) in 11 of those 15
tory to say that investors can earn excess returns by months. Trend following has been a good hedge
both buying when valuations are cheap and buying against long-tail risk, not only in 2008 but also going

Figure 2. Excess Returns for Composite Carry and Trend-Following Strategies in the 15 Worst
Months for Global Equities between 1985 and 2009

Return (%)
20

15

10

–5

–10

–15

–20

–25

–30
8

97
8

90

1
09

01
8

08
9

08
02
/9

/9
0

/0
/0

/8

/0

/
p/

p/

p/

b/

b/

n/

n/
l/
ug

ug

ug
p
ct

ct

Ju
Fe

Fe
Se

Se

Se

Se

Ju
Ja

Ja
O

Composite Carry Composite Trend Following

Note: The composite carry portfolio is composed of four carry strategies in fixed-income and foreign exchange markets, whereas the
composite trend-following portfolio is composed of trend-following strategies in commodity, equity, and fixed-income futures as well
as in foreign exchange.
Sources: Based on data from Bloomberg and author's own calculations in Ilmanen (2011).

©2012 CFA Institute  cfa pubs .org JUNE 2012  59

CFA Institute Conference Proceedings Quarterly

further back in history. It is difficult to tell a risk these assets earn disappointing long-run returns
story that would explain why this strategy has and much lower risk-adjusted returns than low-
delivered positive results. Again, the explanation is volatility or low-beta assets.
probably behavioral. The third reason defensive investments earn
better returns is that managers who have a bench-
Volatility. The fourth strategy is volatility. I
mark treat risk symmetrically. Managers with a
like to think of this strategy in terms of a two-tail
benchmark will increase risk (tracking error)
distribution: The left tail is about buying or sell-
whether they are acting more aggressively or more
ing insurance, and the right tail is about demand
conservatively than the benchmarks because both
for lottery tickets.
actions are deviations. When many investors think
The best-known insurance strategies involve
in those terms, absolute risk may no longer be
various methods of selling equity index volatility.
rewarded in financial markets.
Doing so earns good long-run returns but at the risk
of huge losses when bad times occur. Selling insur- Liquidity. Before the financial crisis, investors
ance pays off for others when the losses happen. In had a complacent sense that they would always be
2008, investors experienced a huge systemic crisis rewarded for bearing illiquidity risk or holding illiq-
that caused many to lose their fortunes. The key idea uid assets. By 2006–2007, however, the various illi-
in financial theory is that investments should earn a
quidity premiums had been beaten down to
positive risk premium if they perform poorly in bad
ridiculously narrow levels, and then an avalanche
times. If certain investments make a little money
of bad events occurred in 2008. Despite this experi-
almost all the time but incur concentrated losses at
ence, evidence remains that investors earn a long-
the worst times, those are the things that, in theory,
run reward for bearing illiquidity.
should offer the largest risk premiums. Investors
may have underestimated the severity of this risk Illiquidity shows up in many different contexts,
before 2008, and in the aftermath of the crisis, few are such as venture capital, commodities, equities, and
willing to sell this type of insurance. government bonds. Unfortunately, no one metric
At the other tail are high-volatility assets, which exists that can measure illiquidity in all of these
I call “lotteries,” that offer speculative returns. It different types of illiquidity. Working with the
turns out that the most volatile assets within every Dutch pension fund company APG, I developed an
asset class offer surprisingly poor long-term illiquidity score for different types of investments.
returns. Meanwhile, low-volatility assets offer sur- As shown in Figure 3, I used a scale of one to five
prisingly good returns for taking a small risk. along the x-axis, with five being the least liquid.
Defensive investments often provide the same or Then I plotted the 20-year average return against
perhaps better absolute returns and certainly much those illiquidity estimates, which shows a positive
better risk-adjusted returns than their more specu- relationship. The results may be overstated because
lative peers. This result explains why low-risk not just total illiquidity but also other risks increase
investing has become popular in stock markets and along the scale, and there are upward biases in
may also attract attention outside of equities. For venture capital, hedge funds, and private equities
example, Panel B in Figure 1 shows that there is no because of voluntary reporting. Nevertheless, over
reward for holding bonds with a maturity greater this time period, there seems to have been a positive
than five years and Panel C shows that CCC rated long-run reward for illiquidity.
bonds underperformed more highly rated credits.
There are three main explanations for the better Conclusion. Style diversification is more effec-
returns on defensive investments. First, people tive than asset class diversification. If investors com-
overpay for insurance. They also overpay for lottery bine various asset classes, they can create a portfolio
tickets. When they do the same thing in financial that is similar to a global market-cap portfolio. They
markets, those types of investments tend to be over- will not get much volatility reduction because the
valued and deliver poor long-run returns. market direction dominates, so their Sharpe ratios
A second reason is that many investors will not will improve only by a small amount. By combining
use leverage, and the strategies with the best reward various trading styles that have, on average, near-
for risk often require it. Many institutions dislike zero pairwise correlation, investors can add good
volatility but dislike leverage even more. Therefore, diversifiers (which may also have attractive Sharpe
these institutions often create a substitute for lever- ratios). With this approach, they can cut their vola-
age by buying the most volatile assets in any asset tility in half and double their Sharpe ratio, but it
class, making this segment overpriced. Predictably, does require shorting and leverage.

60  JUNE 2012 ©2012 CFA Institute  cfa pubs .org

 Understanding Expected Returns

Figure 3. Compounded Annual Return for Various Asset Classes Plotted

on Illiquidity Estimate, 1990–2009
Return (geometric mean)
17

Venture Capital
15
Private
Equity
13 Hedge Fund
Emerging Fixed Income Timber
11 Small Equity
Emerging Equity

9
High Yield
Fund of Funds
U.S. Fixed Global Real Estate
Global Income
7 Infrastructure
Government
Global
Equity
5 Global REITs
Deposit Rate Commodity Futures
3
0 1 2 3 4 5 6
Illiquidity Estimate

Sources: Based on data from Bank of America Merrill Lynch, Barclays Capital, Bloomberg, Cambridge
Associates, Citigroup, FTSE, Global Property Research, Hedge Fund Research, Ibbotson Associates
(Morningstar), Ilmanen (2011), J.P. Morgan, Kenneth French’s website, MIT-CRE, MSCI Barra, NCREIF,
Standard & Poor's, and UBS.

Broadening Perspectives to Time- has many risks, including high concentration and
career risks; being early often equals being wrong.
Varying Expected Returns Besides valuation indicators, investors can
Once institutions address the equity risk in their consider such indicators as measuring the macro
portfolios, they can find other ways to enhance environment and investor risk aversion. In almost
returns at the margin. One way is market timing. It any investment, there are both short-run momen-
should not be a primary form of risk taking, but it tum and long-run reversal effects to consider
can add value. because of speculative dynamics in investor behav-
For years, both academics and practitioners ior. Timing, however, is difficult because system-
were negative about any kind of market-timing allo- atic value signals and discretionary stories tend to
cation, but this attitude is changing because market give opposing messages.
volatility has been so violent over the last 10 years. For example, almost everyone is bearish on
With hindsight (and arguably, with foresight), val- Europe right now, but the valuation indicators tell
investors to buy European assets because they are
uation indicators provided investors with some use-
very cheap compared with those of other regions.
ful contrarian signals.
It will be a few years before it is known which view
There are many ways to practice market timing.
is right.
One approach is contrarian timing based on valua-
In the last few years, important valuation ques-
tion indicators—buying investments that have
tions have been raised about emerging markets ver-
underperformed and selling those that have done sus developed markets, as well as about oil versus
well. Figure 4 shows more than 100 years of other commodities. Investors should consider
forward-looking real yields for various U.S. asset whether there are ongoing structural changes that
classes. Comparing these starting yields with future justify ignoring valuation indicators. These are
returns shows positive correlations. These valua- incredibly difficult questions, so it is no surprise that
tion indicators seem to be able to slightly improve many investors decide to stay with neutral alloca-
long-run Sharpe ratios. Of course, market timing tions over time.

©2012 CFA Institute  cfa pubs .org JUNE 2012  61

CFA Institute Conference Proceedings Quarterly

Figure 4. Forward-Looking Real Yields of Various U.S. Asset Classes,

1900–2010
Yield (%)
20
18
16
14
12
10
8
6
4
2
0
1900 10 20 30 40 50 60 70 80 90 2000 10
Real Baa Credit Real Treasury Rental, 1960–2009
Recession Dummy Smoothed Real Earnings
Sources: Based on data from Bloomberg; Robert Shiller’s website (www.econ.yale.edu/~shiller/);
Ibbotson Associates (Morningstar); Moody’s; Davis, Lehnert, and Martin (2006); Lincoln Institute of
Land Policy; Kozicki and Tinsley (2006); Federal Reserve Bank of Philadelphia; Blue Chip Economic
Indicators; and NBER.

For many institutions, the big question is how 10 years ago emphasized adding illiquidity premi-
to achieve a 4–5 percent real long-run return when ums and alpha through hedge funds. There is some
equities are offering only 4–5 percent returns, fixed- role for all of these approaches.
income returns are averaging 0–2 percent, and cash The way I prefer to improve performance
is giving negative real returns in developed mar- includes working with a range of investment styles.
kets. The first answer is that investors need to adjust Investors can add return by diversifying across mul-
their expectations lower, which is happening. But it tiple premiums, not just across different asset classes.
is natural to try to boost returns beyond these slim After investors harvest long-run rewards from multi-
offerings. Growth-related premiums can be earned ple sources, they can consider some mild tactical tilts
not only by owning equities but also through hold- to exploit time-varying expected returns.
ing some highly correlated assets. The endowment
model that came into vogue after the previous crisis This article qualifies for 0.5 CE credits.

R EFERENCES
Davis, Morris A., Andreas Lehnert, and Robert F. Martin. 2008. Ilmanen, Antti. 2011. Expected Returns: An Investor’s Guide to
“The Rent–Price Ratio for the Aggregate Stock of Owner- Harvesting Market Rewards. Hoboken, NJ: John Wiley & Sons.
Occupied Housing.” Review of Income and Wealth, vol. 54, no. 2 Kozicki, Sharon, and Peter A. Tinsley. 2006. ‘‘Survey-Based Esti-
(June):279–284. mates of the Term Structure of Expected U.S. Inflation.’’ Bank of
Dimson, Elroy, Paul Marsh, and Mike Staunton. 2010. “Credit Canada Working Paper 06-46 (December).
Suisse Global Investment Returns Yearbook 2010.” Credit Suisse Pastor, Lubos, and Robert F. Stambaugh. 2003. “Liquidity Risk
(February). and Expected Stock Returns.” Journal of Political Economy, vol. 111,
Frazzini, Andrea, and Lasse H. Pedersen. 2010. “Betting against no. 3 (June):642–685.
Beta.” NBER Working Paper No. 16601 (December).

62  JUNE 2012 ©2012 CFA Institute  cfa pubs .org

 Q&A: Ilmanen

Question and Answer Session

Antti Ilmanen
Question: How do you mix strategies do not completely off- uid assets, such as small-cap
momentum and contrarian styles? set each other. stocks or emerging markets.
When value and momentum
Ilmanen: Besides thinking strategies are combined, the result Question: Could there be
about expected returns, investors is basically two alpha sources that industry structure effects as inves-
need to think about how to com- may have a clear negative correla- tors copy decisions of major finan-
bine things. I recommend that tion with each other. Even if inves- cial institutions?
investors view their portfolio tors get only a small return gain Ilmanen: For any kind of finan-
diversification in risk terms and from those as a package, it is good. cial idea, being copied can be good
not just in nominal capital alloca- If investors hold value invest- news at first because it will give
tion. If investors do not consider ments, they should consider add- the idea a tailwind. In the long
asset volatilities, then they do not ing a little bit of momentum so run, however, growing popular-
realize that equities are dominat- they do not miss this diversifica- ity will decrease the ex ante profit-
ing performance and risk even in tion opportunity. Likewise, if an ability of any strategy. There is a
a 50/50 portfolio. It helps to think investor tends to be momentum valid concern that ideas lose value
in terms of volatility rather than oriented, then some amount of when they become public—the
dollars. value is a good diversifier. opportunity to profit from them
Next, think about correla- will be reduced. Academics
Question: Can you talk more
tions. It is good to have some neg- would say that if there is a pure
about why momentum works?
atively correlated investments. risk premium behind these ideas,
Investors have been able to enjoy Ilmanen: I think momentum then some positive reward will be
that government bonds and and trend following reflect sustained. If it is pure market inef-
stocks have been negatively corre- behavioral stories. In general, ficiency, investors learn about it
lated for the past 15 years. Most people extrapolate in expecta- and the premium goes away.
investors have both of these, so tions. They also take such actions The momentum effect, for
they have benefited from some as using stop-loss rules or VaR example, has not worked as well
limits that make investors act as in the last 10–15 years than it did
natural diversification. That was
trend followers on the risk man- in the past, but on average and
not the case during the previous
agement side even if they do not globally, it has worked even after
30 years and may not be the case
chase returns. it was known. I think investors
in the future.
The other part of the behav- should expect somewhat lower
Concerning styles, the best ioral explanation is underreac- returns for strategies that are
combination is value and tion. There is quite a lot of more like inefficiencies, but I
momentum—meaning for value, evidence that markets respond think they will not fall all the way
buy something that is cheap, and quite efficiently to news. News to zero. Returns will fluctuate
for momentum, buy recent win- can have a large instantaneous over time, and the strategy will
ners. Value is highly correlated impact on prices. A little bit of again become unpopular, just as
with the losers of the last few information, however, is almost some systematic strategies did in
years, but it is less highly corre- always left on the table. Investors recent years. Paradoxically, the
lated with the losers of the last discover it in many different con- fluctuating doubts about a strat-
few months. In this way, inves- texts, so part of momentum is the egy’s sustainability may sustain
tors can benefit from good diver- delay effect of past news. It tends these premiums in the long run,
sification, although the two to be more pronounced in less liq- even the behavioral ones.

©2012 CFA Institute  cfa pubs .org JUNE 2012  63