place the assets in the account under the direction of a new manager.
If we have a manager of foreign exchange (FX) who is authorized to
hedge all foreign currency exposure that is not otherwise hedged, our
trustee can provide our FX manager a daily list of the total unhedged
dollar exposure of our composite account to each foreign currency.
We can have all of our assets made available to a single securities-lending
agentÔÇ”the most convenient agent often being the master custodian itself.
Each of these management information services comes, of course, with
a price tag. Our adviser can help us select a custodian and also which man-
agement information services would seem worth subscribing to.
Every investment fund should have a custodian, usually a bank.
The custodian safeguards the assets and provides transaction reports
and asset listings showing book values and market values.
The custodian can also be a valuable source of management informa-
tion, such as performance measurement and portfolio composition.
Investment FundÔÇ™s Organization
f I were asked to evaluate the organization of an investment fund, how
I would I go about it?
IÔÇ™d be interested in prior performance, of course, but that is only pro-
logue. Past performance, either absolute or relative to peers, is not by itself
a valid criterion. Good past performance might be the fortuitous result of a
poorly designed investment program, and poor past performance might be
the random result of a well-designed investment program.
The only thing that counts is the future. Hence, our evaluation should
be future oriented. We will want all the quantitative measures that are rele-
vant. But ultimately, an evaluation comes down to asking the right ques-
tions and concluding with qualitative judgments. What are some of the
right questions? My suggestions follow.
Does the fund have a written statement of Investment Policies? Are
they well thought out?
Do all key decision makers buy into these policies, and do they under-
stand their rami´¬ücations?
Do the Investment Policies de´¬üne an overall risk constraint? Why and
how was the measure of risk chosen?
Is the risk constraint appropriate for the fund sponsorÔÇ™s ´¬ünancial
Is the investment return objective de´¬üned as the highest return that can
be achieved within that risk constraint?
128 EVALUATING AN INVESTMENT FUNDÔÇ™S ORGANIZATION
How long have these objectives been in effect?
Net of all costs, how well have these objectives been met?
What is the Policy Asset Allocation?
What was the rationale underlying this asset allocation?
How close is the Policy Asset Allocation to an Ef´¬ücient Frontier?
Were adequate sensitivity tests on risk, return, and correlation assump-
tions carried out in Ef´¬ücient Frontier studies?
How many diverse asset classes were included in the Ef´¬ücient Frontier
How many asset classes is the adviser prepared competently to recom-
mend and manage?
How close has the fundÔÇ™s actual allocation been to its Policy Allocation?
Does the plan diverge from its Policy Asset Allocation tactically? If so,
who decides? With what results?
How does the fund go about rebalancing?
THE FIDUCIARY COMMITTEE
Does the fund have a written statement of Operating Policies?
Who are the members of the fundÔÇ™s ´¬üduciary committee? What are the
criteria for membership, and why?
How much experience do the members have in institutional portfo-
lio investing, and how much time each year do they devote to this
Is the ´¬üduciary committee suf´¬üciently oriented to the long-term, or is it
overly concerned with short-term performance?
What decisions does the committee reserve for itself, and what deci-
sions does it delegate to its adviser? Why?
How much con´¬üdence does the committee have in its adviser?
What constraints does the committee place on its adviserÔÇ”either
through limits on its openness to new ideas or the frequency of its
How does the committee judge the effectiveness of its adviser?
How experienced is the adviser? What is its size, its continuity, and its
commitment to excellence?
How does it go about retaining good people?
How able is it to advise on asset allocation? What are its processes
How able is it to evaluate alternative investments such as those de-
scribed in Chapter 5?
How well-researched and supported are its recommendations (or
How well does it communicate with the investment committee?
How well does it do in the continuous education of committee members?
How does the adviser monitor each of the fundÔÇ™s managers and main-
tain an up-to-date understanding of all the factors that impact the pre-
dictive value of that managerÔÇ™s past performance?
What triggers a recommendation to terminate a manager?
What is the rationale for retaining each of the fundÔÇ™s current
How does the adviser go about ´¬ünding out whether there are better
managers available that it might be using?
How solid are the adviserÔÇ™s administrative support services?
What steps has it taken to mitigate manager risks? What are its au-
dit procedures? How does it control managersÔÇ™ use of derivatives?
How much possibility is there for a manager to penetrate the fundÔÇ™s
In each asset class, was an index fund considered? Why or why not
was an index fund chosen?
What do the adviserÔÇ™s other clients say about their experience with the
adviser? What has been the adviserÔÇ™s client turnover?
130 EVALUATING AN INVESTMENT FUNDÔÇ™S ORGANIZATION
In each asset class, net of all costs, how have the fundÔÇ™s active man-
agersÔÇ”including terminated managersÔÇ”performed relative to an in-
dex fund alternative? How have they performed relative to other
Has the fund stayed within its overall risk constraint?
If the fundÔÇ™s historic risk has been materially below its overall risk con-
straint, could performance have been improved by taking more risk,
closer to the fundÔÇ™s overall risk constraint?
Structure of an Endowment Fund
hat is the purpose of an endowment fund or foundation? Its purpose
W is to throw off a perpetual stream of income to support the operations
of the sponsoring organization. All endowment policies should ´¬‚ow from
Key criteria for endowments:
For planning and budgeting purposes, the stream of income should be
For the health of the organization, the magnitude of that stream of in-
come should, over the long term, maintain its buying power. ÔÇťWhile
´¬üduciary principles generally specify only that the institution preserve
the nominal value of a gift, to provide true permanent support, institu-
tions must maintain the in´¬‚ation-adjusted value of a gift,ÔÇŁ writes
David Swensen of the Yale endowment fund.1
THE TOTAL RETURN, OR IMPUTED INCOME, APPROACH
If your organization is already using the Imputed Income approach to
determine the amount of money to transfer each year from the your en-
dowment fund to your sponsorÔÇ™s operating account, you may care to skip
The traditional approach to income recognition by endowment funds is
David F. Swensen, Pioneering Portfolio Managment (The Free Press, 2000), p. 27.
132 STRUCTURE OF AN ENDOWMENT FUND
materially ´¬‚awed. It de´¬ünes income by the accounting de´¬ünition of divi-
dends, interest, rent, and sometimes net realized capital gains as well. If an
investment approach is kept reasonably intact, the year-to-year stream of
dividends, interest, and rent is quite predictable. But the impact on income
is a key consideration if anyone suggests a major change in the endow-
mentÔÇ™s investment approach. Moreover, if realized capital gains are in-
cluded in the de´¬ünition of income, that creates a wild card.
The trouble with income de´¬üned as just dividends, interest, and rent
is that it impacts investment policy both seriously and detrimentally. To
provide a reasonable amount of income, the sponsor is motivated to in-
vest a major proportion of the endowment in bonds and stocks that pay
high interest or dividends. And if income is running short of what is
needed, the sponsor can adjust the investment portfolio to increase the
dividends and interest.
Such an approach dooms the second criterion, of maintaining the en-
dowmentÔÇ™s buying power over the long term. ThatÔÇ™s because the market
value of bonds, for example, does nothing to maintain buying power (un-
less one considers relatively new in´¬‚ation-linked bonds). There is no con-
cept of real returnÔÇ”investment return in excess of in´¬‚ation.
Because the de´¬ünition of income has such a great impact on investment
policies, the de´¬ünition of income must be revised before we can consider
Fortunately, a concept has been developed that does a pretty good job
of meeting both of the above criteria, and it is widely used today. It is
known as the ÔÇťTotal ReturnÔÇŁ or ÔÇťImputed IncomeÔÇŁ approach.
The Total Return Approach begins with the concept that we should
recognize as income only an amount that might, long-term, be considered
real incomeÔÇ”income in excess of in´¬‚ation. So the ´¬ürst question is: How
much real return can we realistically aspire to earn on our endowment
fund over the long term?
That depends on how we invest the endowment fund. ItÔÇ™s obvious,
then, that the way to maximize the endowmentÔÇ™s real return over the long
term is to maximize its total returnÔÇ”the sum of accounting income plus
capital gains, both realized and unrealized. We go right back to Chapters 3
and 4, ÔÇťInvestment ObjectivesÔÇŁ and ÔÇťAsset Allocation,ÔÇŁ as they are the
LetÔÇ™s say we decide on an asset mix of 80% common stocks, 20% ´¬üxed
income, or 75/25, and after consulting historical returns of those asset
classes, we conclude that we can prudently expect a real return over the
The Total Return, or Imputed Income, Approach
long term of 5% per year. That says we can recognize 5% of market value
as ÔÇťImputed IncomeÔÇŁ each year.
But with this asset allocation the market value will ´¬‚uctuate widely
from year to year, and if Imputed Income equals 5% of that ´¬‚uctuating
value, then the ´¬‚uctuations in annual income may be greater than our
sponsor can live with. Therefore, letÔÇ™s de´¬üne Imputed Income as 5% of a
moving average of market values. While there are many variations, we
have found that a sound de´¬ünition of Imputed Income is: X% of the aver-
age market value of the endowment fund over the last ´¬üve year-ends (ad-
justed for new contributions and adjusted for withdrawals in excess of
Imputed Income, if any).
Appendix 9 at the end of this chapter describes how the Imputed In-
come method works. This approach includes the following advantages:
The sponsor gains a fairly predictable level of annual income, not sub-
ject to large percentage changes from year to year.
The sponsor learns early in the year the amount of income it will with-
draw from the endowment fund. This is a big help in budgeting.
Investment policy cannot be manipulated. The only way our sponsor
can increase Imputed Income is by investing more successfullyÔÇ”by
achieving a higher rate of total return, long term.
What is the downside of the Total Return Approach? There may be in-
tervals when the fundÔÇ™s market value declines for several years in a row,
and the dollar amount of Imputed Income may actually decline. There is
no way to avoid this possibility, since all markets are volatile. But markets,
over a great many years, have eventually bounced back each time. As they
do, the fund, under the Total Return Approach, should regain the buying
power that it lostÔÇ”provided the rate of Imputed Income was established
responsibly in the ´¬ürst place.
Once the sponsor knows the amount of Imputed Income for the cur-
rent year, it can decide when during the year it will withdraw the money. A
withdrawal usually requires a sale of stocks or bonds (or mutual funds),
since cash should not generally be allowed to accumulate. Dividends and
interest should be reinvested promptly.
Warning: Members of the fund sponsor may agitate to raise the Im-
puted Income percentage, and clamor hardest at the wrong time. For ex-
ample, during the 20-year intervals ending in the late 1990s, investment
returns far exceeded the 5% real return assumption that formed the
134 STRUCTURE OF AN ENDOWMENT FUND
philosophical basis for the Imputed Income formula of many endow-
ment funds. At that time, members of certain fund sponsors agitated
successfully to raise the Imputed Income percentage to 51/2% or 6%ÔÇ”
just before the market plunged in 2000ÔÇ“2002. It is easy to forget that a
market debacle like that has occurred before and will occur sometime
again. The higher the market goes, the higher the probability that it will
occur. My advice, therefore is: DonÔÇ™t make a change. If we have estab-
lished our Imputed Income formula soundly in the ´¬ürst place, letÔÇ™s not
tinker with it.
David Swensen of the Yale endowment fund has articulated the danger
well: ÔÇťIncreases in spending soon become part of an institutionÔÇ™s perma-
nent expense base, reducing operational ´¬‚exibility. If the rate of spending
rises in a boom, an institution facing a bust loses the bene´¬üt of a cushion
and gains the burden of a greater budgetary base.ÔÇŁ2
ÔÇťOWNERSÔÇŁ OF THE ENDOWMENT FUND