this route. As discussed above, hedge funds might appear in any of
Berkeley’s four major asset categories—equities, excess returns,
diversifying, and defensive, depending on the strategy of the fund.
A potential advantage of this categorization is the facilitation of
ex-post performance attribution relative to ex-ante assumptions.
Did the fund or private deal live up to its initial promise? On the
other hand, benchmarking and performance attribution may
become more complicated under this unusual breakout.
In the end, we applaud the wide-ranging approaches taken to
defining asset categories by university endowments. We expect that
some of these experiments will be conspicuous successes and lead
to emulation by other investors, some may be applicable to a small
number of specific universities, and others may prove problematic
and transient. Future research in this domain will help speed up
Amid a rapid rise in the range of investment alternatives and
struggles to find associated paradigms for governance and oversight, large university endowments provide a reference point for
best practices. Surveying public disclosures complements the broad
approach of direct surveys like that of NACUBO. Self-reports lack
full transparency and leave much unstated, but investors seeking
to emulate the success of endowments will find useful insights. In
particular, endowment self-reports provide templates for reinvigorating traditional asset categories, key elements of investment processes and reporting alike.
John M. Mulvey, PhD, is a professor of operations research and
financial engineering and a founding member of the Bendheim
Center for Finance at Princeton University. Contact him at
Margaret Holen, PhD, is a private investor and retired partner of
Goldman Sachs, where she held leadership positions in the Strategies
Group responsible for analytics in the firm’s revenue divisions. She
continues to serve as a member of the GS Foundation Investment
The authors gratefully acknowledge the support of First Republic
Bank for the contributions of student research assistants to this work.
1. In research conducted in the third and fourth quarters of 2015, we focused on reports
for the fiscal year ending in 2014, which were the most recent and widely available
at that time. We included reports from neighboring time periods when 2014 data was
unavailable. The fifty reports found were among the largest seventy endowments. We
excluded endowment information provided in audited financial statements of portfolio
holdings absent supporting materials.
2. In a few cases, institutions list differing asset categories and allocations across different publications. In these cases, our summary tables and statistics note the asset
allocation from the annual report.
3. See also, e.g., Financial Times January 25, 2016, “Richest U.S. Universities Reel in
4. “Readers of previous Annual Report letters may observe that [the strategy section] of
the annual letter has remained consistent—and will again this year. This is because
as a long-term investor our investment strategy remains relatively constant while our
implementation is flexible in order to take advantage of capital market opportunities”
5. We found three endowments that did not fit neatly within the dominant norm or novel
trends. Two of these indicate broad categories for “alternatives” (Indiana University
Foundation and University of Kansas Endowment). The third has categories aligned
with the dominant norm but with “inflation hedge” and “deflation hedge” in place
of “real assets” and “fixed income” (University of Oklahoma Foundation), without
6. For example, Princeton discusses an aggregate equity category covering public,
private, direct, and absolute return. Harvard shows breakouts of private equity by
geographic region in its performance discussion, granularity it excludes in its asset
7. The Harvard report that discusses the new framework is from the 2015 year-end; the
allocation included in our survey was from 2014.
8. Three of the five categories in this paradigm would fall into the NACUBO aggregation for alternatives. Four of the endowments, among the thirty that conform to this
paradigm, include an aggregation for alternatives across private equity, real asset,
and absolute return categories (University of California System, University of Illinois,
Wellesley, and University of Rochester).
9. A similar grouping is used by University of Washington, which presents an aggregate
capital appreciation category that aligns with Princeton’s equity aggregation.
UC Berkeley groups all but its defensive categories together under a theme of
10. Though the importance of excellent third-party managers is broadly emphasized by
the reports in our survey, few provide any insight into the selection process.
11. It is unclear from the public reports if the institutions include all commodities-related
equities in this combined category, or, if they do, what adjustments are made to their
equity category benchmarks to reflect that.
12. Duke’s $7-billion endowment is managed by a stand-alone firm, Duke University
Management Company, known as DUMAC, which manages Duke’s assets along with
those of a few associated institutions. Duke reports on its endowment in a separate
dedicated publication, “Duke University’s Endowment Snapshot,” which at two pages
is substantially more compact than Stanford or Princeton’s (Duke 2014). DUMAC
also provides a relatively terse web page titled “Risk Management” (Duke 2016) that
highlights the benefits of diversification and the portfolio’s high Sharpe ratio ( 1. 12 vs.
0.38 for its benchmark over a twenty-five-year period).
13. Like Duke, Harvard presents an equity aggregation that includes public and private
along with a commodities category, though the latter consists solely of direct exposures and was being phased out of the portfolio.
14. This is a common breakdown. For example, the NACUBO–Commonfund survey
(NACUBO 2015) separates core fixed income from distressed debt/high yield.
15. UC Berkeley, technically the endowment of the University of California, Berkeley
Foundation, provides an unusual degree of disclosure. It publishes an annual report
and its investment policy document along with a separate policy overview document
that it describes as a “supporting narrative” for its investment process (BEMCO
2014a, 2014b, 2014c ). Like other institutions discussed here, it has a dedicated
management company, though Berkeley’s endowment is significantly smaller, at
$1.5 billion as of June 30, 2014, and has had a modest return history.
16. Though its asset categories are defined broadly, Berkeley’s annual report notes that
its holdings in the defensive bucket consist exclusively of bonds and cash (BEMCO
2014a). They also set and justify a benchmark portfolio: 83.5 percent global equities
and 17. 5 percent treasury bonds.
17. Another example with some resemblance to this paradigm is Pennsylvania State
University. It presents four aggregates: growth combines private and public equity;
inflation spans TIPS, commodities, natural resources, and real estate (like Texas
Teachers); defensive contains only nominal bonds; and diversifying contains a single
hedge fund category (Pennsylvania State University 2014).
18. UVA’s foundation endowment, with $6.9 billion, is managed by a dedicated company,
the University of Virginia Investment Management Company, or UVIMCO. UVIMCO
reports on endowment management in a twenty-page annual publication, with roughly
fourteen pages devoted to separate chapters on investment strategy, investment
policy, risk management, and investment performance. It is among the most extensive
and substantive discussions in our sample. UVIMCO’s 2015 report includes a page
on operational due diligence and its 2014 report features an interview with the chief
19. In the words of UVIMCO’s Chief Executive Officer Larry Kochard, the organization’s
“core competency [is] selecting exceptional managers” (UVIMCO 2015).