Hewlett Foundation Case Study (2005)

9 – 2 0 5 – 1 2 6 R E V : J A N U A R Y 2 6 , 2 0 0 6

Professor Luis M. Viceira and Research Associate Helen H. Tung prepared this case. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2005 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School.

L U I S M . V I C E I R A

H E L E N H . T U N G

Investment Policy at the Hewlett Foundation (2005)

In early January 2005, Laurance (Laurie) Hoagland Jr., the vice president and chief investment officer of the William and Flora Hewlett Foundation (HF), and his investment team met to finish their recommendations to the HF Investment Committee of a new asset allocation policy for the foundation’s investment portfolio, which was then valued at $6.4 billion. The HF Investment Committee included three members of the foundation’s board and three nonboard members. Hoagland was not a member of the committee.

It was a warm Sunday afternoon in Palo Alto, and sunlight inundated the bright, open spaces of the foundation’s building. The building, completed in May 2002, reflected HF’s commitment to energy conservation and environmental protection. More generally, the building was also a reflection of HF’s approach to philanthropy, based on encouraging innovation and attaining demonstrable results in the solution of complex social and environmental problems. The building was only the fifth in the nation to receive a gold-level certification by the U.S. Green Building Council for its outstanding and innovative “green” design. HF had earned this recognition by addressing a wide range of building-related environmental issues concerning site design, water and energy efficiency, materials and resources, and indoor environmental quality and by creating a healthy and productive workplace.

If the investment team went ahead with the proposal, and the Investment Committee approved it, HF would adopt a new asset allocation policy that included a substantial reduction in the overall exposure of the investment portfolio to domestic public equities (from 30% to 21%) and a significant increase in the allocation to absolute-return strategies (from 10% to 20%) and to TIPS1 (from 7% to 13%). The new asset allocation policy also included a recommendation to combine the increase in the allocation to absolute-return strategies with an “equitization” and “bondization” program. This program would transform the market-neutral exposure of these strategies into exposure to domestic equities, domestic bonds, and TIPS. The overall effect of the changes to the policy portfolio would be a small increase in the Sharpe ratio2 of the entire portfolio of the foundation.

1 U.S. Treasury Inflation Protected Securities. TIPS differ from regular Treasury bonds in that their principal and coupons grow with inflation.

2 The Sharpe ratio is equal to the difference between the expected return on the portfolio and the expected return on cash, divided by the standard deviation of the portfolio return.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

205-126 Investment Policy at the Hewlett Foundation (2005)

2

Hoagland and his team also needed to make a final decision that afternoon on a complementary recommendation to the HF Investment Committee to pledge approximately 5% of the total value of the portfolio to Sirius V, the latest fund at the firm Sirius Investments, which specialized in global distressed real estate investments. HF had a long-established relationship with Sirius. Three years ago, the Investment Committee had pledged 3% of the portfolio value at the time to Sirius IV, which was expected to generate a 20% net return per annum. Hoagland and his team were confident that Sirius could deliver superior returns because the managers at Sirius had “raw talent,” the firm’s business strategy had high barriers to entry for competitors, and its aggregate investment had low downside risk. This confidence was the result of a continuing relation with Sirius and a recent careful evaluation of the firm that included visiting Sirius’s overseas offices.

However, Hoagland was aware that, while it was not unusual for an institutional investor like HF, a private nonprofit foundation, to invest 5% of assets with one management firm in a major asset class like U.S. equities, it was less common to invest 5% of assets in a single manager of a nontraditional asset class such as distressed real estate.

The William and Flora Hewlett Foundation

HF was established in 1966 by William (Bill) R. Hewlett, his wife Flora, and their eldest son, Walter. At year-end 2004, it was among the eight largest U.S. private foundations by asset size, with $6.4 billion. Bill was one of the cofounders of the Hewlett-Packard Corporation (HP). HP was one of the most successful electronics companies in the United States, with a market capitalization at fiscal year-end 2004 of $54 billion, revenues of $80 billion, and 151,000 employees worldwide. A pioneer of California’s Silicon Valley, Bill inspired legions of loyal employees and entrepreneurs through his unassuming approach to business and enthusiasm for engineering. Bill had passed away on January 12, 2001 and had left HF a large bequest of $4.1 billion, consisting primarily of HP and Agilent3 stock.

The grant programs supported by HF reflected Bill’s other life passions and values. The foundation’s resources were focused on six program areas (conflict resolution, education, the environment, performing arts, population, and U.S.-Latin American relations) and on special initiatives in philanthropy, global affairs, and neighborhood improvement (see Exhibit 1a for a description of the programs). Grants awarded by HF were often paid in installments, usually over three years. Thus, they were liabilities for the foundation, in the form of grants payable. In recent years, gifts and grants paid had increased substantially thanks to a significant growth in assets, attributable mainly to Bill’s last bequest (see Exhibit 1b for grants made in 2004). In 2003, the foundation gave 90% more in gifts and grants than in 2000 (see Exhibit 2 for gifts and grants paid from 1998 to 2003). Foundations, unlike universities, relied on neither annual gifts from alumni and friends to grow their assets, nor tuition or government funds supporting research and education, to supplement their operations or distributions. For nonoperating foundations4 like HF, return on invested assets was the only source of income.

HF used a guideline of paying out annually 5.25% of the three-year moving average value of investment assets, which was in line with the spending policy at other large private foundations.5 Private nonoperating foundations like HF were exempt from annual income taxes (federal and state)

3 In 1999, HP spun off Agilent, which was a leading manufacturer of scientific instrument and analysis equipment.

4 A nonoperating private foundation was a private foundation with grant making as its primary activity. It did not operate facilities or institutions devoted to a charitable activity, as did its operating counterpart.

5 The Hewlett Foundation believed that this protected grant recipients from investment return volatility.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

Investment Policy at the Hewlett Foundation (2005) 205-126

3

and local taxes provided they paid annually at least a “distribution amount” in the form of “qualifying distributions.”6 The distribution amount was approximately 5% of the average market value of the foundation’s net investment assets, less a federal excise tax on net investment income.7 Qualifying distributions included mainly grants and gifts as well as administrative expenses such as salary and rent associated with making grants. They excluded the cost of generating investment return. This cost consisted mainly of internal investment staff salaries and overhead as well as manager fees. HF had estimated this cost to be approximately 35 basis points (or 0.35%) of the market value of the assets.8 (Exhibit 2 illustrates calculations of distribution amounts and qualifying distributions for HF from 1998 through 2003.)

If a foundation did not pay the full distribution amount qualifying distributions during the year, it was able to carry forward the difference (if positive) for one year without incurring a tax penalty. After the first year, the foundation was levied an excise tax equal to 15% of the undistributed amount each year the failure was uncorrected. After the fourth year, the tax could be as large as 100% of the undistributed amount.9

Regardless of payout, private nonoperating foundations were required to pay an annual federal excise tax equal to 2% of their net investment income. A foundation qualified for a reduced rate of 1% if it made qualifying distributions for the year equal to or in excess of the average payout for the preceding five years.10 Many of the largest private foundations tried to meet the criteria for the lower tax rate, as HF had from 1998 to 2003, though occasionally a foundation chose to pay the higher tax rate. For example, the Lilly Endowment paid the higher tax rate in 1998, which amounted to nearly $6 million in lost funds for grant seekers.11

Asset Allocation at the Hewlett Foundation

HF investment policy aimed at providing uninterrupted support for worthwhile programs into perpetuity by, at minimum, maintaining or growing current asset size and spending power in real terms. In Hoagland’s own words, “A long-term asset allocation policy should meet our objective, while allowing us to sleep well at night; that is, it should maintain a consistent spending pattern without sharp ups and downs.” HF believed that a first step toward meeting this objective was to hold a well-diversified portfolio that displayed the most effective and efficient trade-off between portfolio return and risk, as measured by the annual standard deviation of the portfolio return.

HF changed its asset allocation policy in response to modified capital market assumptions and new investment opportunities. Changes in the policy required approval of the Investment Committee. This committee also oversaw the selection of benchmarks in each asset class, which were used to evaluate performance.

6 Internal Revenue Service form 990, www.irs.gov.

7 Specifically, the distribution amount was 5% of the fair market value (FMV) of investment assets less excise tax on net investment income less unrelated business tax plus recoveries. The adjustments were minimal, so the distribution amount was effectively 5% of the FMV investment assets. In the case of HF, the distribution amount as a percentage of the FMV investment assets averaged 4.6%–5% from 1998 to 2003.

8 These fees included internal costs and costs of separate accounts with public securities managers.

9 Mary Ann Hakin, “Minimum Distributions From Private Foundations,” Langan Associates, December 11, 2002, p. 9.

10 Ibid., p. 34.

11 Reed Abelson, “Some Foundations Choose to Curb Donations and Pay More Taxes,” The New York Times, February 24, 2000.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

205-126 Investment Policy at the Hewlett Foundation (2005)

4

Changes in the asset allocation policy were infrequent and, when they happened, were typically small. But periodically there were major changes. The last one had happened in October 2001, as a result of an in-depth asset allocation study that Hoagland initiated after joining HF in early 2001. The new policy sharply reduced the allocation to U.S. public equities (from 50% to 30%) and significantly increased the allocation to equities in foreign developed markets, private equity, and real assets. (Exhibit 4 shows HF’s asset allocation policy during recent times.)

The new asset allocation proposal that Hoagland and his team were about to send to the Investment Committee was the result of another in-depth study of HF’s asset allocation. This study included a new assessment of HF’s capital market assumptions, followed by simulations of the long- term performance and risk of different asset allocation policies under the new capital market assumptions. These two pieces constituted a major input in informing HF’s asset allocation decisions. HF’s investment team saw this quantitative analysis as being particularly useful in examining the impact of marginal changes in allocations on the risk-adjusted performance of the portfolio, from one asset class to another.

Another important input in HF’s asset allocation decision was manager selection. HF had a preference for increasing allocations to those asset classes for which it had a comparative advantage in selecting managers. For example, private equity was one of the asset classes with the largest dispersion in performance among managers. HF had traditionally had access to the most recognized private equity managers in the country and intended to keep those relations in the future. Real assets represented another asset class where HF felt it had been able to pick particularly good managers.

Ultimately, the investment team made asset allocation recommendations to the Investment Committee based on a thorough discussion process that might trim extreme asset recommendations coming out of the quantitative analysis or the manager-selection criterion.

Capital Market Assumptions and Asset Allocation Study

HF made long-term (up to 20 years) projections of the expected return, return volatility, and return correlation of all the major asset classes in its portfolio.12 These projections were based on long- run historical estimates as well as expert analysis of prevailing market conditions. The newly assessed capital market assumptions reflected an anticipated environment of lower expected returns resulting from low interest rates, stable macroeconomic conditions (particularly inflation, which HF expected to stay around 2.5% per annum in the foreseeable future), and high valuations of virtually all investment assets. (Exhibit 3a shows the new capital market assumptions, and Exhibit 3b shows the assumptions adopted in October 2001.)

The reduction in the expected return on high-quality fixed-income assets (bonds and TIPS) was the result of lower going-in yields on long-term bonds purchased today and low inflation expectations. Hoagland and his team expected a 4.4% return on domestic bonds, or 20 basis points above the current yield on 10-year Treasury bonds. This difference reflected the fact that they hoped to gain some incremental return from the credit risk embedded in the HF fixed-income portfolio, offset by the fact that they intended to keep the average duration on this portfolio shorter than the duration of the 10-year Treasury bond.

12 Using standard practice in the industry, HF formulated asset-class projections for continuously compounded returns. Thus, the expected-return projection corresponded to a geometric average return. Expected simple return (or arithmetic mean return) projections could be obtained by adding half the return variance projection to the expected continuously compounded return projection.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

Investment Policy at the Hewlett Foundation (2005) 205-126

5

In assessing the expected return on equities, Hoagland and his team gave significant weight to the U.S. historical experience. Over the twentieth century, the premium return on stocks over bond yields had been about 5.8% per annum. This would imply an expected return on equities of about 10.0% per year at current bond yields. But they also considered the analysis of current equity market conditions by some prominent experts who deemed them as rich. Some analysts argued that the U.S economy was now more stable and better managed than during most of the twentieth century. Consequently, one should expect a lower equity premium in the future, perhaps as low as 4%. Other analysts argued for an even lower equity premium, considering that the current valuation of U.S. equities relative to earnings or dividends was high when compared to the historical experience.

To support their view, these analysts noted that the current dividend yield on U.S. stocks was about 2%, and the value of the market was about 24 times the average earnings over the last 10 years. Both measures were well below their historical means of about 4.8% and 17 times, respectively. The expected return on equities was approximately equal to the dividend yield, plus the expected increase in earnings per share, plus the expected change in the price/earnings multiple. Assuming that the dividend yield stayed at 2%, that the earnings multiple remained unchanged, and that earnings per share increased about 4% per year in nominal terms (or about 1.5% per year in real terms), the expected return on equities would be about 6%. This implied a forward-looking premium of about 2% over current yields on long-term bonds. Of course, a contraction in multiple valuations toward their historical average would result in an even lower expected return on equities.

After much debate, Hoagland and his team assumed an expected return on equities of 8%.13 This return was somewhat smaller than the expected return implied by the historical equity risk premium and the current yield on long-term bonds.

Hoagland and his investment team also lowered the expected return on the absolute-return portfolio. They expected that it would return 5.5% per annum going forward. This was 150 basis points lower than the 2001 expected-return assumption and significantly lower than the actual performance of the portfolio since its inception in 1995. (See Exhibit 3a, Exhibit 3b, and Exhibit 9a.) They felt this assumption was justified in light of current market conditions and trends in the investment management industry, where an increasing number of institutional investors were allocating significant fractions of their assets to absolute-return strategies.

Based on the new capital market assumptions, Hoagland and his team had produced a new asset allocation proposal whose most significant changes were a reduced exposure to domestic public equities (from 30% in 2003 to 21%) and an increased allocation to absolute-return strategies (from 10% to 20%) and TIPS (from 7% to 13%). This policy implied a total exposure to asset classes other than cash of 120%, with a negative 20% net exposure to cash. This built-in leverage in the overall portfolio was the result of the complementary bondization and equitization program described below. (Exhibit 4 shows the policy proposal for 2005, and Exhibit 6 compares HF’s actual asset allocation and the proposed asset allocation policy with those of its peers.)

The new asset allocation proposal had been adopted after simulating the long-term performance and risk of different asset allocation policies under the newly assessed capital market assumptions. HF projected the 10-year and 20-year performance results of the proposed policy and a hypothetical conservative allocation policy (shown in Exhibit 5a) using Monte Carlo simulations. These simulations included a spending policy of 5% per annum in real (or inflation-adjusted) terms. 14 The

13 This expected continuously compounded return assumption, together with the assumption about return volatility, implied an expected simple return projection of 9.3% for equities.

14 This was defined as achieving a nominal spending rate of 7.5%�a real spending rate of 5% and an inflation rate of 2.5%.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

205-126 Investment Policy at the Hewlett Foundation (2005)

6

conservative policy had twice the allocation to bonds as the proposed policy and a considerably smaller allocation to equities, which resulted in a lower expected return and a significantly lower volatility. Exhibits 5b through 5e summarize the results from the Monte Carlo simulations. Exhibit 5b shows the probability of occurrence of three events of interest to HF (described below), and Exhibit 5c shows the median portfolio asset values in 2005 dollars (e.g., in inflation-adjusted terms). Exhibit 5d and Exhibit 5e show the probability distribution of asset values for the proposed allocation in 2005 dollars and in nominal terms, respectively.

The Monte Carlo experiment examined the probability of occurrence of three events relevant to HF. The first event of interest to HF was a “seriously negative event,” defined as at least one occurrence of a 20% decline in funds available for grants in inflation-adjusted terms from any previous peak. The second event was an increase in the real (or inflation-adjusted) value of the portfolio.15 The third event of interest was maintaining HF current spending in real terms. The simulations showed that, under the new capital market assumptions, the probability of the seriously negative event was roughly similar for both allocation policies at both horizons, but that increasing the real value of the investment assets and maintaining spending in real terms was much more likely under the proposed policy than under the conservative policy. Moreover, the simulations also showed that the median value of the assets in 2005 dollars would be larger than the starting value of $6 billion under the proposed policy, while it would be considerably lower under the conservative policy.

Based on the results of the Monte Carlo simulations, Hoagland and his team had concluded that the slight protection the conservative asset allocation provided in the event of a “seriously negative event” did not warrant forgoing the benefit of the higher expected return of the proposed asset allocation.

Donor Stock-Sale Program

In 1986 HF adopted a policy of reducing donor stock holdings over time to lower the risk associated with being heavily weighted in a small number of stocks. This policy contrasted with the policy at some other large foundations such as the Packard Foundation, Lilly Endowment, and W.K. Kellogg Foundation, which held a substantial amount of donor stock. For instance, the Packard Foundation investment portfolio, consisting primarily of HP and Agilent stocks, was valued at $13 billion in 2000. By March 2002, it had dropped to $6 billion. This decline in asset value forced the foundation to reduce grants to $250 million in 2002, down from $460 million in 2001, a 46% reduction.16

By November 2000, the percentage of donor stock in the foundation’s portfolio was 2%. However, this percentage increased to 61% in early 2001 with the $4.1 billion bequest of HP and Agilent stock from Bill Hewlett’s estate. In response to this situation, Hoagland and his team conducted a study of the impact of HP and Agilent stock on the overall investment portfolio.

The study analyzed the increase in portfolio return and risk with varying percentage holdings of donor stock. (See Exhibit 3b for the capital market assumptions used in the study.) HF expected the long-term return of the donor stock to be equal to that of the domestic equity market (10.3% at that

15 The starting value used in the simulations was $6 billion.

16 John Boudreau, “Packard, Hewlett Foundations Carry On, Continue Investment Strategies,” The San Jose Mercury News, March 31, 2002.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

Investment Policy at the Hewlett Foundation (2005) 205-126

7

time).17 HF assumed the standard deviation to be 45%, compared with 40% for private equity, the highest-risk asset class in the diversified portfolio. In contrast, the portfolio excluding the donor stock had an expected return of 11.1% and a standard deviation of 14.6% (see the October 2001 policy in Exhibit 4). HF estimated that the correlation of donor stock with the portfolio excluding donor stock was about 30%. Based on these numbers, the study revealed that increasing the percentage of donor stock in the portfolio increased overall portfolio risk but had a negligible impact on expected return. For allocations of 10% and 20%, total portfolio standard deviations were correspondingly 15.1% and 16.5%, with donor stock accounting for an increasing percentage of total portfolio risk. To compensate for the increased portfolio risk, HF required the donor stock to return a premium over domestic equity. However, historical data suggested that this was difficult to achieve. From 1981 to 2000, the historical return on domestic equity had been 13.6% per annum, while the return on HP had been 14.0%. This study led HF to conclude that they should continue their long-standing policy of avoiding high portfolio concentration in a small number of stocks.

Beginning in May 2001, HF systematically sold HP and Agilent shares so that by February 29, 2004, the stocks accounted for only 7% of total portfolio value. (Exhibit 7a shows the number and value of donor shares and their value relative to the total portfolio over this time.) The donor stock- sale program included weekly and discretionary sales, writing covered calls on unsold shares, and writing knock-in forwards on Agilent shares.18 (Details of the stock-sale proceeds are shown in Exhibit 7b.) From June 2002 to June 2003, HF suspended the sale of Agilent stock because of a sharp drop in Agilent stock price.

In 2001, a major event affected HP that did not alter HF’s disciplined sale of HP stock. On September 4, 2001, HP announced the acquisition of Compaq. This was the start of a much publicized argument between the management and some HP shareholders over the desirability of the merger. The Hewlett family’s position was to vote against the merger. The Hewlett and Packard foundations concurred, though all parties arrived at the conclusion independently. As of December 7, 2001, the heirs of both founders (including trusts and foundations) controlled 18% of HP shares, making them the largest block of HP shareholders.19 Despite opposition, the merger was approved on March 20, 2002 by a narrow margin, and the merged company debuted on May 7, 2002. During November and December of 2001, HF sold 5.5 million HP shares, 13% of its total HP holdings.20 It was undisclosed whether the sale was fulfilled through private placement or the open market, and so it was unclear whether HF had any control over the voting of the shares sold.

Investment Process at the Hewlett Foundation

HF set asset allocation policies and managed the donor stock sale internally but used external managers to invest its portfolio funds. In the case of publicly traded equities, 70% of the portfolio

17 This was a conservative assumption, given that both HP and Agilent stocks had historical betas above one. Over the period 1961–2000, HP stock had a beta of 1.4, an average total return (capital gains plus dividends) of 12.6% p.a., and a return standard deviation of 33.5% p.a. A value-weighted basket of all U.S. stocks traded on the NYSE, AMEX, and NASDAQ had an average total return of 10.7% p.a. and a return standard deviation of 15.3% p.a. The source of the data on HP and the U.S. stock market is the Center for Research in Security Prices (CRSP) of the University of Chicago.

18 A knock-in forward contract was a forward contract (on Agilent shares, in this case) that automatically came into existence only when the shares in the spot market traded at a price equal to or larger than a predetermined forward or “instrike” price. Thus, if this event never happened over the predetermined life of the contract, the forward would never become active.

19 Steve Lohr, “He Said. She said. It Just Gets Uglier,” The New York Times, March 17, 2002.

20 Dawn Kawamoto, “Hewlett Foundation Sells Portion of HP Shares,” CNET News, January 4, 2002.

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

205-126 Investment Policy at the Hewlett Foundation (2005)

8

was invested passively in index instruments, and the remaining 30% was invested with managers pursuing active investment strategies. Hoagland headed the investment team at HF, consisting of three directors, an associate, and an analyst. Each director was responsible for an investment category (private equity, real estate, and publicly traded assets and investment research). Hoagland and one of the directors oversaw the absolute-return portfolio, in conjunction with an external manager.

Performance Benchmarks

HF used four methods to evaluate its diversified portfolio performance. The first method was a comparison of the performance of each investment category with that of its chosen benchmark, an index fund whenever possible. This method reflected the investment team’s effectiveness in selecting managers relative to its alternative strategy of passive investment in (mostly) index funds. (See Exhibit 4 for a list of current benchmarks for each asset class.)

The second method measured the team’s asset allocation skill by comparing the performance of an HF “composite benchmark” with a blend of U.S. stocks and bonds with a comparable risk profile, such as a 60%/40% stocks/bonds portfolio. The HF composite benchmark was equal to the sum of the benchmark return on each asset class multiplied by the asset class weight in the policy portfolio.21 The investment team thought that the effectiveness of an asset allocation was difficult to assess in the short run. For instance, portfolios diversified away from U.S. public equities during the bull market of the 1990s experienced lower returns during that period, but recently such portfolios outperformed.

The third method of evaluating HF’s portfolio performance was by comparing its return with those of its peers, other large tax-exempt institutional investors that faced similar investment parameters. The final method was determining whether return on assets had exceeded 5% plus inflation.

Exhibit 8 shows the investment performance of the HF portfolio (excluding donor stock) since December 1996. A dollar invested at the end of 1996 was worth $2.92 (net of management fees) at the end of 2004, while a dollar invested in the composite benchmark was worth $1.86 (net of passive management fees). This translated into an annualized return of 14.3% for the portfolio, versus an 8.1% return for the composite benchmark. This value added was mostly concentrated, but not uniquely so, in the performance of alternative assets (i.e., all assets except domestic fixed income and public equity) relative to their benchmark.

Compensation

Compensation for the directors consisted of a base salary and a bonus based on a quantitative measure and, to a lesser extent, a qualitative evaluation. The quantitative component was based on the performance of the rolling three-year average return of the HF diversified portfolio relative to the HF composite benchmark.

This performance could fall in one of four quartiles. If the HF return was in the first quartile (defined as beating the benchmark by 100 basis points or more), the directors received 150% of the target performance payment, which was itself 20%–40% of the base salary. For second-, third-, and fourth-quartile performance (defined as beating the benchmark by less than 100 basis points, 21 There was also a composite benchmark for each large asset class: public equity (domestic equity, foreign equity, and emerging market equity), alternative assets (private equity, real assets, and absolute return), and fixed income (domestic bonds, TIPS, and high–yield bonds).

For the exclusive use of L. Vera, 2021.

This document is authorized for use only by Lazaro Vera in Global Private Banking Spring 2021 taught by MARCOS KERBEL, Florida International University from Jan 2021 to Jul 2021.

Investment Policy at the Hewlett Foundation (2005) 205-126

9

underperforming it by less than 100 basis points, and underperforming it by more than 100 basis points, respectively), the directors received 100%, 50%, and 0%, respectively, of the target performance payment.

The base salary plus the target performance payment was the competitive market compensation. The qualitative component of the bonus was based on agreed-upon goals outlined during the previous year’s annual review. Unlike compensation at some investment firms, the bonus was not subject to clawback in times of future underperformance.

Absolute-Return Portfolio

HF launched its absolute-return investment program in October 1995. This program sought to invest in strategies that generated market-neutral returns comparable with equity returns, but with substantially lower volatility and little correlation with other asset classes in its portfolio. Hoagland and his team felt that thus far the program had met its objectives. As shown in Exhibit 9a, the annualized return of the HF absolute-return program, from October 1995 to year-end 2003, had outperformed the Russell 3000 Index22 by 230 basis points, while the volatility of the return, as measured by its standard deviation, was commensurate with that of the Lehman Brothers Aggregate Bond Index.23 The HF absolute-return program also performed well relative to the broad U.S. equity market by maintaining a steady 5% return during the stock market decline in 2000 and by minimizing losses to –5% during the global financial market turmoil in the third quarter of 1998.

Several measures indicated that this performance had been largely market neutral with respect to both the equity market and the bond market. The correlation coefficient between the absolute-return program and the Wilshire 5000 Index24 from October 1995 to year-end 2000 was only 0.12. From 2000 to 2003, the program’s beta coefficient relative to the S&P 500 Index was 0.07, and the beta relative to the Lehman Brothers Aggregate Bond Index was 0.01. In terms of risk exposure, the absolute-return portfolio was essentially equivalent to holding cash with some additional volatility.

The absolute-return program, currently valued at $600 million (or about 10% of the current portfolio value), included eight categories. The three largest allocations were in the categories of event arbitrage (36%), distressed investments (25%), and fixed-income and convertible arbitrage (20%). (Exhibit 9b shows a complete breakdown by strategy, and Exhibit 9c includes a detailed description of each strategy.)

Absolute-return investment vehicles, also known as “hedge funds,” were largely unregulated investment pools designed to provide consistent absolute returns by taking maximal advantage of a fund manager’s investment skills while minimizing risk exposure to the overall market. To this end, hedge funds made use routinely of leverage and short-selling when implementing their investment

22 The Russell 3000 Index consisted of the 3,000 largest, by market capitalization, and most liquid stocks based and traded in the U.S. The index accounted for approximately 98% of the total value of all equities traded on U.S. exchanges and was reconstituted annually to include new and growing equities.

23 The Lehman Brothers Aggregate Bond Index was a benchmark index comprising the Lehman Brothers Government/Corporate Bond Index, Mortgage-Backed Securities Index, and Asset Backed Securities Index, which included securities of investment-grade or better quality that had at least one year to maturity and had an outstanding par value of at least $100 million.

24 The Wilshire 5000 Index was considered the total market index comprising stock of companies that met three criteria: they were headquartered in the U.S., their stock was actively traded on a U.S. exchange, and their stock had widely available price information.

For the exclusive use of L. Vera, 2021.