Investment Performance
The year ended June 30, 2009, was a difficult one for investors generally and particularly for strategies like those followed by DUMAC, LLC. During this time period, global equities declined 29.3 percent, real estate declined 40.6 percent, and commodities were down 59.7 percent.* Many of the strategies that contributed to DUMAC outperforming the benchmarks by a sizable amount over the past decade hurt performance in 2008-09 and contributed to a disappointing one-year outcome. Nonetheless, DUMAC's multi-year returns remain strong on a relative and absolute basis.
DUMAC managers seek to achieve an annualized real rate of return of at least 5.5 percent net of fees to fund the university's spending rate and to allow for growth of the endowment after considering the effects of inflation. This long-term benchmark is applied over rolling ten-year periods, rather than year by year, due to the volatility implicit in short-term periods.
As illustrated in the chart below, net returns have consistently been above the 5.5 percent real rate. As a result, Duke was able to meet its spending needs while growing the size of the endowment in real terms.
DUMAC's primary long-term benchmark is the composite of 70 percent Russell 3000 Index and 30 percent BC Aggregate Index (70/30). The Russell 3000 Index represents the broad domestic equity market while the BC Aggregate Index represents the domestic bond market. This mix between equity and bonds was chosen because, historically, a 70 percent exposure to equity investments has achieved the university's 5.5 percent real annual return objective over the long term.
DUMAC aims to exceed the performance of the 70/30 over rolling five-year measuring periods, after deducting manager fees. This relatively long five-year time horizon is used to account for a more complete market cycle. As illustrated in the chart below, DUMAC's diversification into alternative assets has enabled DUMAC to outperform the 70/30 by a wide margin.
DUMAC uses two benchmarks to track its endowment performance relative to its peers: the Cambridge Associates Universe and the NACUBO Endowment Universe. DUMAC seeks to achieve performance over successive rolling three-year periods in excess of the median return for the Cambridge Associates Universe and in the top quartile for the NACUBO Endowment Universe.
The benchmarks differ because Cambridge data is gathered from a group of peer colleges and universities with endowment pools that are similar to Duke's, often managed by professional internal staff. The NACUBO data comes from a broader group, comprising more than 600 colleges and universities, many of which are relatively small and may lack investment staff resources.
Comparative data for the year ending June 30, 2009, is not yet available. For the three-year period ending June 30, 2008, Duke ranked 2nd out of 50 institutions reporting in the Cambridge Associates Universe and 3rd out of 699 institutions reporting in the NACUBO Endowment Universe.
DUMAC's investment policy identifies target allocations in each asset class as well as permissible levels of variation from these targets. Over a one-year horizon, DUMAC measures its actual allocation of assets against the target allocations. In periods like the year ended June 30, 2009, when nearly all asset classes had negative returns, the policy benchmark is particularly tough given its inclusion of absolute return indices (e.g. inflation plus 5.5 percent).
* MSCI All Country World, NAREIT, and GSCI