1

The Ford Foundation: A Case Study

April 2004

The Ford Foundation, while not the largest U.S. foundation, is the largest one established before the current generation, and has thus been a precedent-setting institution in many ways. With some $10.5 billion in assets, and 14 foreign offices (all in developing countries) as well as a headquarters in New York City, it funds programs in asset building and community development; education, media, knowledge, and religion; and peace and social justice. As a result of a 1969 act of Congress, the Foundation – like all private foundations – is required to pay out 5% of the current value of its portfolio each year, plus investment expenses that average about 0.3%. Administrative expenses (such as staff salaries) not related to the Foundation’s investment activities are counted toward the 5% payout requirement. The Foundation also has a stated goal of maintaining the purchasing power of its portfolio, so its investment objective is to earn 5.3% plus the rate of inflation. The Foundation received its last donation from the Ford (Motor Company) family in 1947 and does not expect to receive any new donations. It must therefore earn all of its income from investments. Linda Strumpf, the Foundation’s chief investment officer, has led the institution’s investment division since 1992. She was formerly an equity fund manager, at the Ford Foundation and previously at a commercial investment management firm. Strumpf is noted not just for her investment achievements but for her exotic adventures, including diving in the ocean to feed sharks and having once been charged by an elephant in Africa. As she gazes at the original Picasso drawing installed in her office as part of the organization’s noted art program, she ponders the following questions: Asset allocation and risk budgeting

How much risk (in the sense of fluctuation of asset values) should the Foundation take? What mix of asset classes best balances the desire to avoid risk against the goal of earning 5.3% plus the rate of inflation? How does one determine this? Should the Foundation adopt this mix as a “policy portfolio,” and automatically rebalance to it, or should the mix just be used as a benchmark, with discretion used in determining the actual asset mix? What is the long-run expected return, from this point forward, on U.S. equities? International equities? Nominal bonds? Inflation-indexed bonds (TIPS)? Cash? What are the risks of each of these asset classes?

2

Should the Foundation use “alternative” asset classes, such as hedge funds, private equity, real estate, and commodities to enhance the likelihood of earning 5.3% plus the rate of inflation? Given the recent popularity of hedge funds, which typically (but not always) “hedge out” market risk and attempt to earn returns purely from manager skill, what kinds of long-run returns should we expect from these funds? How risky are they?

Implementation questions What blend of active managers and index funds is appropriate? Should this blend vary by asset class? Stated another way, how much active risk should the Foundation take? How should active managers be selected for a given asset class? How many active managers are required to adequately diversify active risk with an asset class? Should this number vary by asset class? How should the Foundation decide when to “fire” a management firm? Should the Foundation select equity managers by “style” (growth, value, small capitalization, etc.) or should it employ generalists whose benchmark is the broad market? How should the Foundation’s investment staff be evaluated for compensation and performance review purposes?

The remainder of this case is background information provided for those who would help Linda Strumpf resolve these dilemmas. Recent asset allocation and performance data for the Ford Foundation are in the Appendix.

- - -

Principal asset classes used in endowment and foundation investment management Overview The principal asset classes in which endowed institutions invest are stocks (equities), bonds, and cash. In addition, “alternative” investments have recently taken on a major role in the management of

3

endowed institutions. Alternative investments include not only hedge funds, which are currently popular, but also private equity, real estate, commodities, and other specialized strategies. We devote a subsection to them. Equity investments Because they provide the expectation of a higher return than bonds or cash, stocks constitute some 60% to 70% of the investment assets of U.S. institutions including foundations, endowments, and pension plans. The risk and return characteristics of stocks are reviewed later, and compared with those of other assets. The principal categories of stocks in which institutions typically invest are:

• U.S. stocks • International developed-country stocks (Western Europe, Japan, Canada, Australia, and a few

other countries) • Emerging market stocks

U.S. stocks typically form the bulk of the equity portfolios of U.S. institutions, with non-U.S. stocks held in smaller proportions (up to 25% of the total invested in stocks). International developed-country stocks are an important component of most investors’ portfolios because they tend to diversify the risk of investing in just one country (the United States) and because they expose the investor to unique companies and industries that are only available in non-U.S. markets. While U.S. stocks have outperformed non-U.S. stocks in the past decade, there have been long historical periods when international stocks beat U.S. issues. This occurred over almost the entire period from 1975 to 1989. Emerging market stocks are riskier than the other categories but are believed by many investors to have brighter growth prospects; in U.S. institutions, they typically constitute 5% or less of the total invested in stocks. Active management versus index funds. An index fund invests in each stock in an index (say, the S&P 500) in an amount proportionate to the stock’s weight in the index. No attempt is made to beat the return on the index. This approach is sometimes called passive investing. Management fees are low, but any hope of earning a return higher than that of the index must be abandoned. Index funds are available for every major category of liquid asset, including U.S. stocks, international stocks, and bonds. Index-fund investors reason that on average across all managers, the investment management industry cannot possibly beat the indices because they sum to the market. In other words, active management is a zero-sum game relative to the index; one manager’s outperformance must come at the expense of other managers’ underperformance. Therefore, index-fund investors argue, it is better (and cheaper from a fee standpoint) simply to hold the index.

4

The majority of stock funds, however, are actively managed. Investors who participate in these funds reason that many managers outperform the indices for long periods, and that a relatively modest rate of outperformance – say, 1% per year – accumulates to a large difference in wealth over long periods. (A $1 million investment invested at an 9% annual rate of return grows to $5.7 million in twenty years, but at 8% it grows to only $4.6 million. The difference is more than the original investment.) By careful selection of managers and styles, these investors argue, institutions can achieve better-than-average results. The fact that some 30% of all institutional equities are indexed, and 70% are actively managed, indicates both methods are widely accepted. Investment “style.” Many U.S. equity portfolios, and some non-U.S. portfolios, are concentrated in one size category (large, medium-sized, or small companies) and/or in one methodology for picking stocks (“growth” or “value”).1 Some institutions build a diversified fund of stocks by combining portfolios managed according to several different styles. Alternatively, the institution can select one or more “core” equity managers whose approach includes elements of both growth and value investing, and companies of different sizes. A core manager can be either indexed or active. Fixed-income investments Like stocks, fixed-income investments (bonds and cash) come in many categories:

• U.S. investment-grade bonds (Treasuries, mortgages and corporates ) • U.S. high-yield bonds • Non-U.S. bonds • Inflation-linked bonds (principally U.S. Treasury-issued TIPS) • Emerging-market debt • Cash (Treasury bills and commercial paper)

For most institutions, bonds denominated in U.S. dollars form all or most of the fixed-income investment program. Investment management firms offer funds in each of the categories named. Within each category, the most important characteristic of a fund is its duration, or interest-rate risk. While pension funds, which have long-duration liabilities that are fixed in nominal terms, often invest in long-duration portfolios of bonds, most endowments and foundations find that funds with an intermediate duration (3 to 5 years) offer the best mix of risk and return characteristics.

1 Growth investors seek companies that are experiencing strong increases in sales and profits, and do not mind paying higher-than-average multiples (the stock price exp ressed as a multiple of the current year’s profits) for the possibility of participating in the company’s future prosperity. Value investors, representing an almost diametrically opposed point of view, seek companies that are selling for less than their fair value – that is, they hope to pay lower-than-average multiples – and expect to benefit from price increases when markets recognize the fair values. Unlike growth investors, value investors may buy into stable or declining companies. Many investment managers combine aspects of growth and value investing and cannot be neatly categorized into one group or the other.

5

Many U.S. institutions hold high-yield bonds as well as investment-grade bonds. High-yield bonds can be quite volatile. To control the amount of credit risk in the portfolio, the allocation to high-yield bonds is typically limited. Inflation-linked bonds are of special interest to institutions, such as foundations, that have liabilities that grow with inflation. The inflation-linked bonds issued by the U.S. Treasury have no default risk. They are called Treasury Inflation-Protected Securities or TIPS. Both the principal amount and interest payments on these bonds are linked to changes in the consumer price level, so there is no possibility of the bond’s return falling short of inflation if held to maturity. The total return on these bonds consists of the current interest rate on these bonds (about 2% for the longest-term issues) plus the inflation rate. Cash is an important reserve asset for endowed institutions and similar organizations. Obligations need to be settled in cash, and it is not always practical to liquidate stocks or bonds when cash is needed. Some investors maintain that cash is “idle” and not “invested,” but as long as it is earning interest, this is not a valid argument. Alternative investments Many investors, particularly university endowments, foundations, and wealthy families, have recently turned to so-called alternative investments. These investments are said to offer attractive returns with little or no exposure to, or correlation with, traditional stock and bond markets. The principal categories of alternative investments are shown in Exhibit 1, classified into illiquid investments (in which money is locked up for a period of time) and liquid investments (in which the investor can get his or her money out relatively quickly).

Exhibit 1 Types of Alternative Investments

Illiquid alternatives Liquid alternatives Private equity Commodities

– Venture capital Hedge funds – Buyouts (U.S. and international) – Market-neutral equities

Real estate properties2 – Long-short equities Timber – Fixed income arbitrage Oil and gas partnerships – Event or merger arbitrage – Capital structure arbitrage – Global macro – Distressed securities

2 For the sake of this discussion, real estate equities (REITs) are considered equities, not alternatives. However, some investors consider REITs to be alternative investments.

6

Alternative investments typically appeal to investors who (1) believe the stock and bond markets are overpriced or offer too low an expected return, (2) want to earn an “absolute return,” statistically uncorrelated with traditional assets, or (3) believe they can outperform traditional investments by identifying alternative managers with unusual skill in selecting securities (and especially in selling them short). All of these reasons are intertwined. The track record of these alternative investments in performing as advertised is mixed, and success in this area relies on selection of skilled managers. These investments are appropriate only for large institutions with a strong staff, and with an investment committee that understands that the uncorrelated behavior and inflation hedging properties are not guaranteed. Returns and risks of principal asset classes While most readers are familiar with the relation between risk and expected return, it pays to review briefly the historical evidence that this relation exists, and to consider the possibility that while risk will almost certainly continue to be rewarded in the long run, that reward may be somewhat lower in the future than it has been in the past. Historical returns on stocks, bonds, and cash Stocks, which fluctuate more than fixed-income investments, need to offer a risk premium (called the equity risk premium) or higher expected return in order to be attractive to investors. The size of this risk premium going forward is a topic of great contention. Historically, the S&P 500 index has provided investors with a total return (capital gain plus dividends) of about 10.4% per year over the period from 1926 (when accurate records began to be kept) through 2003. The seemingly small difference between this rate of return and the lower return on bonds (5.4%) or cash (3.8%) accumulates to a huge difference in total wealth when compounded over a long period. Exhibit 2 indicates that a hypothetical investment of one dollar in the stock index in 1926 grew to $2,285 by 2000, while bonds grew to only $60, and cash provided an even lower return that barely outpaced inflation. (Exhibit 3 provides summary statistics, including standard deviation – a measure of risk – for the data used to construct Exhibit 2.)

7

0.1

1

10

100

1000

10000

1925

1929

1932

1935

1938

1941

1944

1947

1950

1953

1956

1959

1962

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

2003

Stocks $2,285

Bonds $60

Cash $18

Inflation $10

Exhibit 2Historical Returns on Principal U.S. Asset Classes, 1925-2003

Source: Ibbotson Associates, Inc., Chicago. Used by permission.

Gro

wth

of $

1.00

inve

sted

12/

31/1

925

8

Exhibit 3

Summary Statistics of Annual Returns on Principal U.S. Asset Classes, 1926-2003

Compound Annual Standard Deviation

Asset Class Return (Risk)

S&P 500 Stocks 10.4 % 19.5 % Intermediate-Term Bonds 5.4 % 4.4 % Treasury Bills 3.8 % 0.9 % Inflation 3.0 % 1.9 %

Source: Ibbotson Associates, Inc., Chicago. Reprinted by permission.

Future prospects of stocks

Some investors assume that past returns are the best guide to future returns, but this assumption is not necessarily valid. The expected return on stocks, and on other assets, is affected by their price. Since stocks performed exceptionally well over 1982-2000, this raised the historical average return, and thus the forecast return if one assumes that the future will be equal to the past. But one might instead, and more reasonably, conclude that future returns will be lower than in the past, because past returns elevated stock prices to a level somewhat above their long-term average. Researchers and practitioners sometimes use a “supply model” to make forecasts that include the effect of the current price. The supply model is an alternative to the future-equals-past model, and proceeds from the assumption that corporate earnings cannot grow faster than GDP (at least not for very long). The dividend discount model, applied to a stock index, then gives the expected return as the dividend yield (which is currently about 1.8%) plus the long-term expected nominal growth rate of GDP.

Risks of stocks and other asset classes Why not hold just stocks? The reason is that stocks are too risky, even over quite long periods. Looking at Exhibit 2, an investor who bought at the 1929 high would have had to wait until 1945 just to break even – and that is true only if dividends were plowed back into the market, rather than being spent. If dividends are not included, then the investor would have had to wait until 1954 to break even. Obviously such a run of poor returns would sink most institutions if they had any spending obligations at all. A safety net consisting of some proportion of assets held in bonds and/or cash would have greatly ameliorated the outcome. The crash of 1929 and subsequent Great Depression might appear to be a one-time event that we do not have to worry about in the future, but the market declines of the 1970’s and early 1980’s were almost as severe in terms of their impact on investors. (The “real” economy, in the sense of production

9

and employment, did not sink nearly as low in the latter episode.) The decline from the market peak in March 2000 to the bottom in October 2002 was 49% and thus compares with the Seventies. The inescapable conclusion is that stocks are risky, even for investors who can afford to hold on for a decade or more. It may superficially appear from Exhibit 2 that the risk of stocks disappears if the investor holds on for the very long run – say, 30 years or more. After all, every old high was eventually surpassed, no matter how far the market fell, or how long the market stayed down. However, no institution with a mandated spending requirement can afford to ignore fluctuations in portfolio values for decades-long periods. During the down part of such period, spending would erode so much of the real value of the institution’s portfolio that there would not be much left to participate in the subsequent recovery. Risks of bonds. Bonds are typically less risky than stocks but are not riskless, because bond prices move opposite their yields (which can rise). The risk of bonds caused by interest-rate fluctuations is illustrated in Exhibit 2, where the line representing bond performance is not smooth. (Intermediate-term bonds, with an average of five years to maturity, are used to represent bond performance in the exhibit.) There are many periods of small losses. However, in the long run, bond investors have faced much less risk than stock investors, as indicated by the relative smoothness of the bond and stock lines in Exhibit 2, and by the standard deviations (a measure of risk) in Exhibit 3. The risk of a bond is proportionate to its duration. Thus, long-term bonds have more interest-rate risk than shorter-term bonds. Risks of cash. Cash and very short-term bonds do not experience these market-value fluctuations due to interest rate changes, so they are viewed as less risky by investors with a short time horizon. Over longer time horizons, however, they are somewhat risky because the interest rates (not the market values) fluctuate, and may earn a return lower than the inflation rate over long periods, as they did in the 1970’s. This causes a decline in the real value, or purchasing power, of the funds invested in cash.

Summary We have seen that stocks have historically outperformed all other major asset classes, and can be expected to continue to do so – but they also have more risk than other assets. Despite their risk, stocks are the only major asset class that has the potential for substantial capital growth over time. This is because participating in the growth and profitability of businesses is inherently more rewarding than merely lending money (which is what bond and cash investors do). Some stock enthusiasts summarize this concept in rhyme by saying, “It is better to be an owner than a ‘loaner.’” This is correct to the extent that one can tolerate the ups and downs of being an owner. The institutional investor’s task is to find the proper balance between owning and “loaning,” between the potential for growth and the need for safety. We now turn our attention to achieving this balance.

10

Selecting an asset mix for an investment fund

Our discussion assumes that the reader is familiar with the basic principle of portfolio diversification. Those who want a refresher can find it in an article on the Ford Foundation web site on which this case is based. Please refer to http://www.fordfound.org/publications/recent_articles/investman.cfm, pages 14-16. By far the most important asset allocation decision is how much to hold in equities (that is, in all categories of stocks combined, including real estate equity and private equity if any) and how much in fixed income (that is, all forms of bonds and cash combined). After that decision has been made, one can then allocate within asset classes. Equities versus fixed income The investor should not rely on any single approach to arrive at the “right” proportion in equities. We recommend combining:

• Risk targeting, and • Peer group analysis.

Risk-target method The risk-target method sets the equity/fixed income allocation according to the amount of risk the institution can tolerate. Exhibit 4 shows the risk of a variety of asset mixes, using realistic assumptions. In the first line of the exhibit, risk is expressed as the probability of losing 10% or more in one year. In the second line of the exhibit, risk is defined as the probability, over five years, of beating a benchmark or hypothetical asset returning the inflation rate plus 5% annually. Note that risk, as defined in the first line, increases as the proportion in equities rises, because equities are more volatile than bonds; in contrast, when risk takes on the definition in the second line, risk decreases as the proportion in equities rises, because equities have higher average returns. (Note that at today’s low expected returns on all assets, even the riskiest mixes have a modest probability of achieving the desired inflation-plus-5% return, before costs.)

Exhibit 4 How much in equities? Risk target approach

Percent in Equities 0 10 20 30 40 50 60 70 80 90 100

11

Probability ≈0 ≈0 .01 .02 .03 .04 .05 .06 .07 .08 .11 of losing >10% in 1 year Probability .11 .16 .21 .25 .29 .33 .37 .41 .45 .47 .50 of beating “inflation + 5%” over 5 years

Based on the risk estimates in Exhibit 4, or on other risk estimates that the institution considers relevant, the chief investment officer or investment committee must decide how much risk can be tolerated. In making this decision, the institution should take into account the investment time horizon, and should moreover keep in mind that riskier mixes are likely to earn higher rates of return over the long run – but that the higher returns are not guaranteed. Time horizon. One of the most important factors in determining how much risk can be tolerated is the time period over which investments are expected to be held. Common sense dictates that an individual investor saving to make a required tax payment next April 15 should invest more conservatively than the same investor saving for a distant retirement. Likewise, when an institution is investing to meet a precisely defined short-term need, the investment approach should be more conservative than when the endowment is intended to be perpetual. It is difficult to quantify the impact that time horizon should have on the risk-tolerance decision; some impressionistic guidelines follow. Assets that are intended to be spent in a year or two should be invested only in cash and short-term bonds. As the investment horizon lengthens beyond two years, there should be a gradual increase in the risk tolerance so that intermediate-term bonds, equities in small amounts, and finally equities in larger amounts are admitted to the mix. Assets that are intended to be held for 20 years or more can be considered perpetual; for reference, the average allocation to equities in perpetual funds in the U.S. is a little over 60%.

Peer groups While pension funds have carefully specified liabilities that serve as a guide to investing the assets, endowed institutions do not have a liability defined independently of the asset value. Thus the endowment’s chief investment officer cannot look to the liability for guidance on what asset mix to hold. As a result it has become customary – and we will argue that it is also reasonable – to rely on one’s peer group. A wide variety of different kinds of endowed institutions face similar investment challenges. These institutions have devoted great effort over a long time period to understanding the issues that bear on the asset allocation decision. Thus there is much information embedded in what other institutions are doing. (Data are available on the allocation, to equities and other asset classes, of leading endowed institutions.)

12

This does not mean that the institution should simply imitate the peer group, however. The task incumbent on the chief investment officer using this method is to ask: “What is different about my situation?” Differences among endowed institutions include those of time horizon, risk tolerance, expenses, and receipts of new money (from donations, or from operations such as tuition collected by a university). Based on an assessment of the organization’s needs and preferences, the chief investment officer and the investment committee of the institution’s Board of Trustees should come to a decision on how its equity allocation should differ (up or down) from the then-prevailing practice as revealed by survey data. Some readers may accuse us of encouraging uncreative “herd” behavior when we suggest keeping an eye on one’s peer group. However, it is only by this means that one can avail himself or herself of the collective wisdom of experienced board, investment committee, and staff members trained in the profession of investment management. Asset allocation within equities Once the equity/fixed income mix has been established, the investment committee must determine the proportions in U.S., developed international, and emerging markets. While the market capitalization of the U.S. is only about half the world total, almost all U.S.-based investors have 60% or more of their equities in U.S. stocks, and a majority has more than 80%. Most U.S. investors, however, should have some international exposure because:

• Diversification (risk-reduction) benefits can be obtained by investing in more than one country • Business cycles outside the U.S. can be out of phase with the U.S. cycle • Restructuring of international companies can provide growth opportunities • Unique companies exist outside the U.S. (international managers often focus on these when

investing money for U.S. clients) There is no simple formula for deciding the percentage in international stocks. Each institution must make this decision. Asset allocation within fixed income Most institutions can afford to take the very slight risk of holding a fixed-income portfolio that does not just contain U.S. Treasury bonds, which are the safest and which consequently have the lowest yields. Corporate and mortgage bonds should also be in the mix. Typically, a single manager or mutual fund is employed to invest assets in all these fixed-income sectors. If the institution chooses to participate in other sectors (high yield, international, etc.), another manager or fund may need to be selected. As noted earlier (but this bears repeating), an intermediate duration should usually be targeted for the fixed-income fund, even in a perpetual institution. The selected manager should receive these instructions from the chief investment officer or investment committee.

13

For the purposes of this discussion, the allocation to cash has been counted within fixed income. Some institutions will want to treat cash as a separate asset class and designate a separate manager (typically the institution’s bank or a money market fund). If cash is treated purely as a source of liquidity for spending, it is sufficient to keep a few months’ expenses in cash. A larger allocation would represent a strategic decision to hold cash for its return and risk-reduction characteristics. Changes in the asset mix Once the asset mix has been determined, the chief investment officer or investment committee needs to monitor the mix on at least an annual basis, and preferably quarterly. Changes in the mix can be motivated by:

• Market movement • Drawdowns (spending) or new money (e.g. donations) • New risk/return/correlation estimates • Changing institutional needs and preferences (e.g. spending rate, risk tolerance, list of

investments permitted by the Board of Trustees) • New asset classes in the marketplace

Some observers believe that the portfolio should be periodically rebalanced to the policy mix. Others prefer to use judgment, rebalancing when it seems prudent to do so, and letting winners ride at other times. An institution should have a policy stating whether or not rebalancing is mandated, and how often. If rebalancing is not mandated, then a permissible allocation range should be established for each asset class. (An example would be 50% to 70% in stocks,) In addition, if there is a change in institutional needs and preferences, such as the spending rate, time horizon, or risk tolerance, an asset shift may be required. Institutions should revisit their asset-allocation policy every two years to incorporate into their decision-making any changes in needs and preferences that may have occurred. Finally, as market conditions change, the asset mix may also need to be changed; if, for example, Treasury-bond yields rose to 7% with little or no increase in inflation, bonds would be much more attractive. One could invest mainly in bonds and still meet the 5.3% spending requirement. This does not mean that the institution should invest only (or mainly) in bonds, since stocks would presumably have correspondingly higher returns under such conditions. It only means that the spending requirement could be satisfied using bonds. The introduction of new asset classes should also be an occasion to revisit the asset-allocation decision. An example is the introduction of inflation-linked bonds by the U.S. Treasury in 1997. These bonds gave U.S. investors an opportunity, for the first time, to buy a default-free domestic instrument that hedges against inflation while simultaneously providing a modest level of interest income. Because inflation is one of the chief enemies of endowed institutions’ continued prosperity, some of these

14

institutions took this opportunity to re-examine their asset mix and move some money from the traditional bond market to this new type of bond. Manager selection

Fund structure Before selecting specific managers for each asset class, the chief investment officer or investment committee needs to decide:

• Whether to use active managers or index funds • Whether to use “core” managers or “style-specific” managers • How many managers

The following discussion focuses on equities. Active management versus indexing The conceptual arguments regarding active management and indexing were presented earlier. On the practical side, indexing is easier for the investor undertaking a manager search because all of the index-fund managers in a given asset class earn almost exactly the same return – the return on the index. Thus the search process becomes a matter of deciding which index-fund manager offers the best customer-support services and charges the most reasonable fees. The search for active managers must, of course, include judgment as to which manager (or pool of managers) is most likely to outperform the index while taking a reasonable level of active-management risk.3 The active-versus-index fund decision can, of course, differ from one asset class to another. For example, some investors index the U.S. large-stock fund, while hiring active managers for international and small stocks, because (these investors believe) the latter markets are priced less efficiently, making it easier for managers to beat the index in those markets. Core versus style-specific managers Some institutions invest in each major “style” of equities – large growth, small value, and so forth – as well as in a large number of other types of funds. The argument for such a complex structure is that active managers are more likely to be successful if they are specialized. This structure can be rewarding to investors who are able to pick successful active managers from the large population of manager candidates. To most investors, it is just expensive and time-consuming. Style balance is desirable, but it

3 For an extensive discussion of issues in selecting a “portfolio” of managers, see Waring, M. Barton, and Laurence B. Siegel, “The Dimensions of Active Management,” Journal of Portfolio Management, Spring 2003.

15

can be achieved through index funds or core active managers as well as through style-focused managers. How many managers? The question of how many managers to hire is related to the previous two questions. One can build a completely style-diversified portfolio using just one manager – a broad-capitalization index fund. However, institutions that seek to add alpha would be well advised to diversify their alpha sources. In addition, more managers are required if one is investing in several different “styles.” This advice needs to be balanced against the cost and staff effort involved in hiring a large number of managers.

Evaluating and hiring managers The steps involved in selecting a manager for a given asset class are:

• Compile list of candidates with suitable profile • Reduce list to 2 or 3 finalists • Interview finalists • Make selection • Negotiate fees (to the extent possible) and conclude manager agreement • Transfer funds

A detailed discussion of each of these procedural steps is in the article on the Ford Foundation website referred to earlier. Briefly, the factors that most institutions consider important in winnowing a list of candidates are the three “P’s” – people, process, and performance – as well as:

• Size of the management firm and the specific portfolio in which investment is being contemplated;

• Fees, which are discussed in more detail below; • Ancillary services, such as custody, performance measurement, advice, etc.; • Type of organization (independent investment advisory firm, broker-affiliated, insurance

company-affiliated, etc.)

Fees Fees are usually calculated as a percentage of assets under management, subject to a dollar minimum. All other things being equal, lower fees are better. An investor seeking strictly to minimize fees should employ index funds. If active management is desired, fees will be substantially higher, but the fund has at least a reasonable chance of outperforming the index. Active management results are not guaranteed – if the fund performs poorly, fees will be charged anyway! Fees are usually negotiable (except with mutual funds), and nonprofit organizations should specifically ask for an eleemosynary discount and insist on paying no more than any other nonprofit client of comparable size.

16

A performance-based fee, or sliding scale wherein the firm participates in the profits of the investor, is sometimes offered to very large investors. Such an arrangement appeals to some investors because (as managers describe it) the interests of the investor and manager are aligned. We would point out that the manager is presumably doing the best he can, even for a flat fee; he or she cannot put the “bad stocks” in the flat-fee account while saving the “good stocks” for the performance account, because he or she doesn’t know which stocks are good and which are bad. A performance-based fee does, however, have the advantage that the investor does not have to pay much when performance is poor. Benchmarking and monitoring Defining performance benchmarks For the performance evaluation process to be fair to both manager and investor (institution), an appropriate benchmark should be defined in advance. The benchmark should be representative of the asset class or style in which the manager is hired. For example, a manager hired to invest in the broad U.S. equity market (including small and mid-sized as well as large-capitalization stocks) should probably be benchmarked to the Russell 3000 or Wilshire 5000 index rather than the S&P 500. A value manager should be benchmarked to a value index (say, the Russell 3000 value index), rather than to an unstylized index. This avoids attributing to the manager any performance that comes from the investor’s choice of style.

Measuring performance and monitoring managers

For actively managed accounts, performance consistently at or above the benchmark is to be hoped for, but no one can outperform all the time. Some investors impose a discipline on manager retention, terminating managers who have underperformed over a market cycle (usually 3, 4, or 5 years). While this is a reasonable approach, the cost of transition to a new manager should be weighed against the cost of keeping an existing, underperforming one. (Transition costs have been estimated at 1% to 2% of capital for core U.S. equity funds and twice that range for international and small-stock funds; the number should be lower for a bond fund.) Therefore, while we would not necessarily terminate a manager for poor performance, the following conditions should raise a “red flag” and would usually result in a termination:

• High turnover in senior investment staff • Major change in investment process • Problems with Securities and Exchange Commission or other legal authorities • Change in ownership

For index fund accounts, the goal is close tracking of the benchmark and low costs. An S&P 500 index fund should track within 0.1% per year. Funds indexed to other benchmarks, especially international and small-stock benchmarks, tend to have more tracking error due to the difficulty of remaining fully

17

invested in all the securities in the index. As with active managers, index funds should be monitored for the kinds of organizational problems referred to above.

Evaluating total fund returns

Endowed institutions need to monitor themselves as well as their managers, because they are responsible for the asset allocation decision. This is true even where a balanced manager is hired. A periodic check is needed to make sure the basic goals of the investment program are being met. These may include earning a return equal to the institution’s required return as determined by spending policy; capital growth in real terms; and avoiding excessive risk or volatility. However, the institution should be reminded that it is to some extent at the mercy of the markets, and that no one with market exposure has good results in absolute terms when all asset classes are down. It is tempting for any institution or individual to cut back on equities, and other risk-bearing investments, when these have not performed well, and to increase them when they have been rewarding. Investors should resist these temptations because they can prove very costly. The best policy is generally to rebalance to the policy mix, which means increasing the equity weight when markets have fallen, and decreasing the equity weight after a rise.

18

APPENDIX FORD FOUNDATION ASSET ALLOCATION AND PERFORMANCE DATA

Asset PercentValues of

($ Millions) Total

EQUITIES AND EQUITY-LIKE SECURITIES

Domestic $4,287.7 41.0% International 1,786.9 17.1 Total Public Equities 6,074.6 58.1

Real Estate 46.7 0.5 Private Equity 875.2 8.4 Total Equities 6,996.5 67.0

FIXED INCOME

Domestic 2,924.6 28.0 International 182.4 1.7 Cash and Accrued Income 344.6 3.3 Total Fixed Income 3,451.6 33.0

TOTAL $10,448.1 100.0%

ASSET ALLOCATION BY SECURITY TYPE

DECEMBER 31, 2003

19

INVESTMENT PERFORMANCE

TOTAL FUND RETURNS

Compound Annual ReturnsFor Periods EndedDecember 31, 2003

One Three Five TenYear Years Years Years

TOTAL FUND 18.7% -3.0% 5.9% 10.7%

Inflation (CPI) + 5.3% 7.3 7.3 7.8 7.8

Total Fund Return in excess of "Inflation + 5.3%" +11.4 -10.3 -1.9 +2.9

Policy Benchmark Return* 19.6% 0.5% 2.5% 8.4%

Total Fund Return minus Benchmark Return -0.9 -3.5 +3.4 +2.3

Northern Trust Universe >$1 Billion Median 24.3% 2.4% 5.0% 9.6%

ASSET CLASS RETURNS

TOTAL EQUITIES 31.2% -1.5% 1.4% 10.2%

U.S. Equities 26.9 -3.0 -0.3 11.1 S&P 500 28.7 -4.1 -0.6 11.1

International Equities 42.6 2.4 6.6 6.9 MSCI EAFE Index 38.6 -2.9 -0.1 4.5

PRIVATE EQUITY -4.4 -23.8 17.7 21.3

TOTAL FIXED INCOME 7.1 8.2 6.6 6.3 Citigroup Broad Interm. & Inflation-Linked Blend 4.8 7.9 7.0 6.9

*Policy Benchmark: S&P 500 45%, EAFE 15%, Citigroup Broad Interm. Term Index 35%, Cash 5%.

20

INVESTMENT PERFORMANCE

Compound Annual Returnsfor Periods Ended

Portfolio Decmeber 31, 2003Values One Three Five Ten

12/31/2003 Year Years Years YearsU.S. Equities

Internally Managed Equity Fund $3,083.7 25.0% -5.8% -1.6% 11.4%

S&P 500 28.7 -4.1 -0.6 11.1

Large Cap Value Manager 416.2 32.1 6.1 - -

S&P 500/Barra Value 31.8 -2.7 2.0 10.6

All Cap Value Enhanced Index Fund 388.3 31.5 3.1 2.5 -

Wilshire All Value Index 32.6 3.1 2.5 10.2

Small & Mid-Cap Value/Core Manager 334.4 34.2 16.2 17.6 15.7

Russell 2500 Index 45.5 6.6 9.4 11.7

Russell 2500 Value Index 44.9 12.8 11.9 13.6

International Equities

European Growth Manager 209.0 32.1 -4.2 1.7 -

U.S.-based Growth Manager 469.9 38.3 -1.3 5.8 -

U.K.-based Mid-Large Cap Value Mgr 467.8 49.9 8.0 10.5 -

U.S.-based Value Manager 522.5 42.7 3.7 6.9 8.9

MSCI EAFE Index 38.6 -2.9 -0.1 4.5

Emerging Markets Manager 174.0 51.5 9.6 10.1 3.5

MSCI Emerging Markets Index 56.3 12.8 10.6 0.2

Real Estate

REIT Manager 152.2 38.2 15.6 16.0 -

Real Estate Partnerships 53.5 27.9 16.8 14.0 11.7

Private Equity 976.0 -4.4 -23.8 17.7 21.3

21

INVESTMENT PERFORMANCE

Compound Annual Returnsfor Periods Ended

Portfolio December 31, 2003Values One Three Five Ten

12/31/2003 Year Years Years YearsFixed Income

Internally Managed Treasury & Mortgage Fund $1,515.8 2.4% 6.5% 6.1% 5.7%Citigroup Broad Interm. Ex. Corps. 2.8 6.7 6.4 6.6

Internally Managed TIPS Fund 499.5 10.1 10.4 - -

Citigroup Inflation-Linked Index 8.4 11.0 - -

Internally Managed Non-U.S. Bond Fund 184.2 15.8 10.4 6.4 -Citigroup WGBI Ex.U.S. 50% Hedged 10.0 8.4 5.4 -

Corporate/High Yield Bond Manager 693.2 13.6 10.4 8.1 -Merrill Corp./High Yield Blend 17.2 10.0 6.4 7.4

Mortgage-Backed Securities Manager #1 201.2 - - - - Lehman MBS Fixed Rate Index - - - -Mortgage-Backed Securities Manager #2 101.7 - - - -

Lehman U.S. Aggregate Index - - - -

Liquid Reserve Fund 5.0 1.8 3.0 4.2 4.9

TOTAL FUND $10,448.1 18.7 -3.0 5.9 10.7