then behave almost exactly like an S&P 500 index fund. An S&P 500 index
fund is a portfolio invested exactly like the S&P 500 index.
A basis point equals 0.01%.
36 RISK, RETURN, AND CORRELATION
valuable tools in managing a portfolio. They enable us to reduce risk by
hedging out risks we donā™t want. Through futures, forwards, or options,
we can choose to reduce our currency risk, or interest-rate risk, or stock-
They also can allow us to take more risk if we like.
They can also be big cost savers. For example, if we want to invest in
an S&P 500 index fund, the purchase of S&P 500 index futures to overlay
a portfolio of cash equivalents may be cheaper (and at least as effective)
compared with buying all 500 stocks for our own account. Buying the
stocks would entail transaction costs, custodial costs, dividend reinvest-
ment costs, and proxy-voting costs. When itā™s time to sell, futures are far
less cumbersome and costly to sell.
As Nobel laureate Merton Miller has said, āIndex futures have been
so successful because they are so cheap and efļ¬cient a way for in-
stitutional investors to adjust their portfolio proportions. As compared
to adjusting the proportions by buying or selling the stocks one by
one and buying or selling T-bills, it is cheaper to use futures by a factor
So why all the fuss about derivatives?
First of all, derivatives can be complex, particularly specially tailored
derivatives that are not exchange-traded. Many people have purchased de-
rivatives without fully understanding all the speciļ¬c risks involved and
have gotten burned badly.
Other investors, through derivatives, have quietly altered their fundā™s
risk/return position substantially without letting their constituents know
until suddenly a blowup has occurred. Recent changes in accounting
rules have helped to lessen this risk through a requirement of sunlightā”
The sheer complexity of certain derivativesā”such as those involving
options, whose pattern of returns is highly asymmetricā”might at times
make it difļ¬cult for some plan sponsors to assess very accurately their full
exposure to the various markets.
Journal of Applied Corporate Finance.
Derivativesā”A Boon or a Different Four-Letter Word?
A 1994 article in Moodyā™s did a good job of summarizing the
The ļ¬nancial roadside is littered with the wreckage of poorly run
derivatives operations. . . . Even entities with excellent internal
controls are not immune from such surprises. . . . Because risk po-
sitions can be radically changed in a matter of seconds, derivatives
activity has increased the potential for surprise. . . .
[But] derivatives often get a bad rap. A frequent message we
hear is that anyone who is involved in derivatives transactions is
tempting fate, and that sooner or later major losses will be suf-
fered as derivatives positions inevitably go wrong. Such messages
Properly used, derivatives have been and will continue to be a
source of risk reduction and enhanced investment performance
for many participants. Therefore, any manager who is not look-
ing at how derivatives can be employed to manage ļ¬nancial and
economic risks, or to enhance yields, is doing his or her investors
As committee members, do we have to understand all this about deriv-
atives? No, we donā™t. Our adviser should understand it. But our adviser
should tell us if a manager uses derivatives, and he should explain in words
satisfactory to us:
What derivatives the manager uses and why;
Losses that could result from use of those derivatives in a volatile, illiq-
uid market; and
The managerā™s approach to risk control.
Although asset classes that use derivatives (such as some arbitrage pro-
grams) require more investor skill to enter, those asset classes are well
worth considering by investment funds. Where an investor can ļ¬nd compe-
tent managers and reasonable terms, use of these asset classes can reduce
the aggregate volatility of an overall portfolio and also increase its overall
expected return. But it pays to be thoughtful about our exposure and to
ask good questions.
38 RISK, RETURN, AND CORRELATION
All rates of return should be based on market values. What counts is
total return, the sum of dividends, interest, and changes in market
value (capital gains or losses).
There are many kinds of risk. Over long intervals there is one measure
that encompasses most of them. Thatā™s volatilityā”how much market
values go up and down over time.
We need to be fully aware of the risks in our portfolio, but we should
not be traumatized by them. The challenge is to mitigate risks by diver-
sifying our portfolio among assets whose returns are not highly corre-
lated with one another.
Setting Investment Policies
nce we have an adviser, our ļ¬rst task as an investment committee is to
O set down in writing our investment policies. Our policy statement
should include our investment objectives, how we will go about investing,
and how we will keep score. We should articulate our principles in a way
that will serve as criteria against which to evaluate both current investment
actions and future proposals.
Well, what is the purpose of any investment fund? For an endowment
fund or a foundation, itā™s to provide a reliable and hopefully increasing in-
come to the fundā™s sponsor. For a pension fund, itā™s to earn money to pay
pension beneļ¬ts. In either case, we should be striving for the highest long-
term rate of return that we can achieveā”within whatever risk limits we be-
lieve are appropriate, of course.
I should emphasize: This is a long-term game. Good returns over
short-term intervals arenā™t very important except as they contribute to
the long-term rate of return. Itā™s the long-term annual rate of return that
At the end of every game, itā™s easy to ļ¬gure out how weā™ve doneā”what
our long-term rate of return was. But no one gets the beneļ¬t of 20/20 hind-
sight when strategizing how to play the game. The only thing that counts is
tomorrow, and tomorrow is an unknownā”anything can happen. So how
do we go about deciding how to invest our money today?
To establish our investment objectives, we must begin by deciding on
three interrelated elements as they apply to us:
40 SETTING INVESTMENT POLICIES
But these are backward. First, we should decide our time horizonā”the num-
ber of years until we need to use our money. That determines how much risk
we can take with our investments. If we need our money tomorrow, we canā™t
afford any risk. Such money shouldnā™t be in an endowment fund.
TIME HORIZON, RISK, AND RETURN
There are major reasons why investing money for an investment fund is
dramatically different from investing oneā™s personal assetsā”other than the
fact that such institutional funds are taxfree. Endowment funds, for exam-
ple, have a perpetual life, while most of the payouts by typical pension
funds are well more than 10 years out.
If I am investing for my family, I must invest essentially for my spouse,
myself, and our children. I really donā™t know when I will need my savings,
and how much I will need, so I must invest conservatively, to prepare for
An investment fund usually knows about what these future payments
must be, and the investment fund should make the most of this critical ad-
vantage. An investment fund should therefore invest with a very long time
horizon, focusing on rates of return over intervals of 10 or 20 years. Why
is this an advantage? Notice from Figure 3.1 how the uncertainty of re-
turns narrows with time like a funnel. Figure 3.1 depicts annual rates of re-
turn since 1926 on large U.S. stocks for intervals ranging from 1 to 20
years. It shows that one-year returns on common stock have been almost
totally unpredictable. Two-thirds of the time, one-year returns have ranged
between +35% and negative 9%. But for 10-year intervals, this span of an-
nual returns has narrowed to a range of +18% to +4%, and the range has
narrowed further for longer intervals.
Clearly, an investment fund should go for the beneļ¬ts of being very
long-term oriented. But too much volatility will make annual payments to
the sponsor too unreliable. This leads to the second element in an invest-
Risk is a hard thing to deal with in setting the investment objectives of a
fund. To quantify our sponsorā™s risk tolerance, our investment committee
Time Horizon, Risk, and Return
S&P 500 Total Return
Annual Rates of Return
Average Returns + or ā“ One Standard Deviation
1 Year 5 Years 10 Years 15 Years 20
All Intervals, 1926ā“2003
FIGURE 3.1 S&P 500 Total Return for All Intervals, 1926ā“2003
Source of Data: Ibbotson Associates, 2004 Yearbook, Chicago, 2004.
could establish the maximum standard deviation of annual returns that we
are willing to incur. But thatā™s tough to specify.
The lower the volatility we can withstand, the lower the long-term rate
of return we can rationally aspire to achieve, and vice versa. But what does
it mean to say, āWe would be willing to incur a standard deviation of X
percentage pointsā? If the ļ¬nancial markets are very placid, weā™ll have no
trouble staying within a standard deviation of X percentage points. But if
the markets are turbulent, a standard deviation of X percentage points will
be a pipe dream. We have no control over future ļ¬nancial markets.
Hence, about the only risk measure thatā™s pragmatic is a relative risk
measure. For example, we could say, āWe can withstand the volatility of
the U.S. stock market as measured by the S&P 500, but not higher.ā His-
torically, the annual standard deviation of the S&P 500 has been about 17
percentage points. What we must wrestle with is this question: Can we
stand the downside volatility of the S&P 500 when it is two or three stan-
dard deviations worse than ānormalā?
More concretely, could we stand a 1973ā“1974 or 2000ā“2002 decline,
when an S&P 500 index fund would have lost nearly 40% of its value? Or
42 SETTING INVESTMENT POLICIES
might our committee lose its nerve at the bottom, decide it should never
have been in such a risky investment program, and sell out at precisely the
wrong time? At the end of 1974 or 2002, such a change of direction would
have been a disaster, as the market regained in the next two years all it had
lost in 1973ā“1974 and was up sharply in 2003ā“2004.
Thatā™s the kind of question we should ask ourselves. Conceptually, an
investment fund should be able to withstand that level of volatility, but
from a pragmatic standpoint, can our sponsorā™s board of directors, or its
future board, withstand it?
That is why, as a benchmark for our total fund, we might establish a hy-
pothetical portfolio of index fundsā”a āBenchmark Portfolio.ā Our objective
would be to incur volatility not greater than that of our Benchmark Portfolio.
How could we know whether a particular Benchmark Portfolio was
appropriate for us? We might see what the volatility of the Benchmark
Portfolio would have been over various long intervals of years and see
whether we, our committee, and our sponsoring organization could stand
that level of volatility.
I think many institutions set their volatility constraint too low relative
to the time horizon that they should establish. As Jack Bogle, founder of
the well-known Vanguard Group, has said, āOne point of added volatility
is meaningless, while one point of added return is priceless.ā
Our return objective is simple. We want the highest long-term rate of re-
turn we can achieve without exceeding our risk constraint.
POLICY ASSET ALLOCATION
Well, where does our Benchmark Portfolio come from? It is a way to mea-
sure the index returns on our Policy Asset Allocation, which we must decide
on ļ¬rst. Our Policy Asset Allocation deļ¬nes the percentage of our portfolio
that we shall target for each asset class. In the next chapter, we shall talk
more about how we might go about deciding on our Policy Asset Allocation.
The Benchmark Portfolio, by measuring the volatility of our Policy
Asset Allocation, quantiļ¬es the maximum risk we are willing to incur.
Does it also deļ¬ne our return objective as well?
One of the good things about a Benchmark Portfolio is that it sets a
Policy Asset Allocation
relative objective, not an absolute objective. An absolute objective would
be something like, āWe want to earn 10% per year.ā Over an extremely
long term, like 20 or 30 years, an absolute objectiveā”especially in real
terms (net of inļ¬‚ation)ā”might be an appropriate objective. For intervals of
fewer years, however, relative objectives are more appropriate, because we
are all prisoners of the market.
But using the return on our Benchmark Portfolio as our investment re-
turn objective still seems inadequate. I think a more appropriate objective
would be āto earn the highest possible rate of return without incurring
more risk than the risk of our Benchmark Portfolio.ā We should be greedy,
aim for the best possible returnā”as long as we stay within our risk con-
straint (set by our Benchmark Portfolio).
We can achieve a rate of return equal to our Benchmark Portfolio if we
invest in index funds identical to that Portfolio. Therefore index returns on
our Benchmark Portfolio should be the minimum return we should aspire
to earn long-term.
We suggested earlier that diversiļ¬cation can help us get more bang for
each point of our portfolioā™s volatility. We should therefore build this di-
versiļ¬cation into our Policy Asset Allocation and Benchmark Portfolio. We
should include any asset class we believe will improve our portfolioā™s ag-
gregate return without increasing its aggregate volatility beyond the limit
we believe is acceptable.
We should review our Policy Asset Allocation periodically for appro-
priateness, but we should change it only with compelling reason. Theoreti-
cally, of course, the best results would come from reducing our allocation
to stocks before stocks enter a bear market and increasing the allocation
before stocks enter a bull market. There are few if any professional in-
vestors who have been able to do this successfully over time, and probably
most would have been better off if they hadnā™t tried.1
Therefore, letā™s not try to time the market. Letā™s try to maintain our
Policy Asset Allocation over the long term. Our Policy Asset Allocationā”
and our Benchmark Portfolio, which measures that allocationā”should be
quite stable over time.
Fidelity Management Company has placed market timing in good perspective with
its āLouie the Loserā illustration. Louie invested consistently the same amount of
money every year for 20 years, 1978ā“1997ā”but unfortunately, always when the
market hit its high for the year. He still had a compound annual return over the 20
years of 15.7%. By comparison, if he had invested each year when the market had
hit its low for the year, his compound annual return would have been only 1.5
44 SETTING INVESTMENT POLICIES
BENCHMARKS FOR MARKETABLE SECURITIES
For each asset class of marketable securities, we should use an index
as its benchmark. After all, if we have no rational expectation to ex-
ceed the index return of an asset class, we should invest in an index
fund for that asset class.
The traditional index used as a benchmark for U.S. stocks is the
S&P 500. This is not an adequate benchmark for our U.S. stocks. The
S&P 500 is essentially a large-stock index with a growth-stock bias.
Even though it measures about 81% of the market capitalization of
all stocks traded in the U.S., our feet should be held to the ļ¬re of all
marketable stocks in the U.S.
Perhaps the best measure of all U.S. stocks is the Russell 3000 in-
dex. The Russell 3000, like the S&P 500, is a capitalization-weighted
index,2 but it consists of the 3,000 largest stocks in the United States
and measures more than 98% of the market capitalization of all U.S.
stocks. It is preferable, in my opinion, to treat large and small U.S.
stocks as two separate asset classes, using the Russell 1000 for large
and the Russell 2000 for small.3
But U.S. stocks account for only about half the capitalization of
all marketable stocks in the world. Therefore, for the purpose of di-
versiļ¬cation, we should include non-U.S. stocks separately in our
The most widely used index of non-U.S. stocks is the Morgan
Stanley Capital International (MSCI) index for Europe, Australia,
and the Far East (EAFE). This also is an inadequate benchmark for
our portfolio. It fails to include Canadian stocks, smaller stocks, or
those of the emerging marketsā”Latin America, much of Asia, east-
ern Europe, and Africa. A better benchmark is the MSCI All Country
Index, ex. U.S. Or, better yet, I suggest treating emerging markets
stocks as a separate asset class, in which case we might use the MSCI
World Index, ex. U.S. tocover common stocks of the developed mar-
kets and then use the MSCI Emerging Markets Free Index for the
For ļ¬xed income, the broadest index of investment-grade U.S.
bonds has long been the Lehman Aggregate Bond Index.
Policy Asset Allocation
BENCHMARKS FOR MARKETABLE SECURITIES (Continued)
As an investment committee member, am I expected to know
about the various indexes that might be used as benchmarks? No, but
a little knowledge about the main indexes will allow us to ask useful
questions of our adviser.
Once we establish our Benchmark Portfolio, we can see how the
performance of our portfolio compares. If, over intervals of three to
ļ¬ve years our performance does not at least equal that of our Bench-
mark Portfolio, we should be asking hard questions as to why.
A cap-weighted index weights each stock in direct proportion to its capital-