Harvard Management Company (Hmc)

1) The Role and Structure of Endowment: As of June 2000, the endowment managed by HMC totaled approximately $18.2 billion. The annual spending from the endowment represented approximately 27% of the total budget of the university. In fiscal year 2000, total endowment spending by the schools was $556 million (or 4% of the value of the fund at the end of previous fiscal year). Each year the University's operational governing body considers the overall financial situation of the University and decides the dividend amount to schools.

* The Average Spending Ratio from the Endowment is 4.6% (low 3.3% and high 5.6%)

* Desired Real Return for the Endowment is 6-6.5%.

* Annual Gifts to endowment is 1.5 % of the fund.

Asset Allocation Policy: Policy Portfolio is constructed by the management and then approved by the Board. Policy Portfolio is constructed according to long-term return and risk assumptions. However, portfolio managers have got flexibility in the short term to decide the asset allocation.

Risk Control: Risk Control was being made by stress testing method.

The Optimal Portfolio Allocation: HMC uses optimal portfolio allocation while constructing their portfolios. They are trying to find the best combination of risky assets to be mixed with safe assets to form the complete portfolio. The advantages of this approach for HMC are:

* It is based on the real return (relative to CPI), risk (standard deviations), and correlations of twelve assets.

* It uses an optimization algorithm which specifies the "efficient frontier" of possible asset combinations according to set of portfolios that would provide the maximum expected return for a given level of risk.

* It includes some untraditional asset classes which could give better return than the traditional ones.

Besides the advantages of this approach there are also some disadvantages which are:

* It is very tied to the current portfolio, allowing only plus or minus 10% deviation.

* HMC made assumptions on 12 asset classes in order to build the portfolio and thus no one would feel confident about all of the assumptions. Since your results will change for the optimal portfolio allocation depending on the model inputs, the estimates of the HMC staff must be reliable and accurate.

* The traditional risk-minimizing attitude of HMC could not be valid in the coming years, because expenses had increased dramatically.

2) HMC develops its capital market assumptions based on historical data from the last 10 years (on a quarterly basis). They rely on the SD of the market and forecast their own return assumption taking into consideration volatility and inflation rate as well.

They focus on real return to see the picture. Also, regarding U.S. and foreign equity their expectations about the market return in the future is lower than their historical averages. But for commodities and emerging markets they expect higher returns for the future than the averages. These expectations are mostly depend on the macroeconomic estimates.

a. Harvard's estimates make sense, because they are forecasting the behavior of the market from their traditionally conservative point of view. They develop their strategy based on the regression equation of the Single-Index Model which takes into consideration not only the historical data but also the future expectation of the market.

3) By adding TIPS, HMC can reduce risk on the portfolio return because the goal of the HMC to ensure a consistent real return. According to this goal, since TIPS are issued by U.S. Treasury and assumed that there is no default risk, by definition TIPS become risk-free assets. Another advantage of TIPS is that they are exempt from state and local taxes. Furthermore, other institutions similar to HMC have already realized that it is worthy to include TIPS in their portfolio.

4) To obtain Exhibit 12 we have copied the data from Exhibit 11 and calculated Mean, Variance, Standard Deviation and Sharpe Ratio. It is important to mention that we restrained all the asset classes by their lower level (greater than 0 for all except Cash, that should be grater than -50%). Then, to generate the efficient frontier we have generated 29 different portfolios by constraining the mean in steps of .25%, as can be seen in Annex 1.

The tangency portfolio was obtained by removing the constraint in the mean.

Finally, the risk-free rate was considered 3.6% in accordance with the information from TIPS.

5) Now we constraint the entire asset classes by their lower and higher permitted values (according to Exhibit 13). We created 9 portfolios, as can be seen in Graph 1.