In a recent article for The Journal of Investment Consulting, John Mulvey and Margaret Holen look at the practice of asset allocation among large U.S. university endowments. They focus on asset category definitions in the hope of throwing some light on “the movement to define asset categories with reference to their target performance or underlying return drivers rather than traditional investment vehicle-types.”
Mulvey is a professor of operations research and financial engineering and one of the founders of the Bendheim Center for Finance at Princeton University. Holen is a retired partner of Goldman Sachs, and continues to serve as a member of the GS Foundation Investment Committee.
In the introduction to their paper, Mulvey and Holen observe that college and university endowments have led the way ahead of other institutions in the move into hedge funds and private equity funds. In 2000, the average endowment portfolio was 23% in alternative assets; in 2014, the corresponding number was 51%. Among endowments with more than $1 billion in assets under management, that number was 57%.
The Abandoned Traditional Model
Of even broader importance, the success of endowments during the period when they were executing that shift has popularized an “endowment model” of institutional investing, which is at odds with the traditional model. In the traditional model, as Mulvey and Holen describe it, managers decided upon allocations in five steps:
1. They defined asset categories (often equity, debt, cash, and real estate) and benchmarks;
2. They conducted asset-liability studies to target allocation levels for each;
3. Then selected investments and implemented;
4. They monitored and reported on the portfolio with respect to performance attribution, concentration risks, and diversification benefits;
5. They periodically assessed current market conditions and considered tactical drifts from their strategic targets within their guidelines.
The endowment model on the other hand, “emphasizes manager selection and opportunistic investing over the top-down decision making incorporated in asset allocations.”
The Current Norm
Mulvey and Holen tell us that their survey of endowments has yielded a composite “current norm” for such institutions. Endowments now have five core asset categories: public equity, private equity, real estate, absolute return, and fixed income.
These categories can be subdivided but the particulars of the subdivision are fluid. There is often a geographical breakdown of public equity, but never (among the institutions surveyed by these authors) a geographical breakdown of private equity, and only rarely one for fixed income.
Private equity is commonly split into venture versus buyout. Fixed income is split into credit versus government bonds.
The “absolute return” category, too, is sometimes subdivided. More on that below.
Positive Trend toward Granularity
Yale describes the core holdings in its absolute return bucket as long/short equity strategies, including both event and value driven sub-classes.
Harvard, on the other hand, says that its absolute return portfolio doesn’t include any equity long/short managers. It adds that it does not view absolute return as “a levered bet on equities.”
Dartmouth’s approach is aligned with Yale. Its one absolute return classification goes by the name “marketable alternative equity.” The stated objective of that category is to generate “equity-like returns with reduced volatility.”
Diversity in practice has driven greater granularity in the definition of asset classes. Ten of the funds in the survey separate their long/short strategies from their other absolute return strategies, in part at least because they expect the L/S equity investments will have relatively high correlation with public equity markets.
Duke has no absolute return category under any name, but it does have a commodities category that includes “commodities related equities as well as private investments in energy, power, infrastructure, and timber.”
UC Berkeley Foundation
Divergences from the current norm, much more so from the long-abandoned traditional model, continue and grow in significance. The general direction is toward functionality.
For example, the UC Berkeley Foundation identifies and categorizes thus: equities; excess return; diversifying defensive. Equity assets are that that are “heavily tied to equity markets and are expected to generate equity-like returns and volatility.” This category includes 37.5% of the portfolio.
The excess return category includes assets “intended to meaningfully outperform equity markets over the long term with less emphasis on interim volatility or liquidity.” This includes PE, VC, and certain real assets strategies.
The other real asset strategies, along with hedge funds, go into the “diversifying” class for Berkeley, because they are intended to generate equity-like returns but to remain uncorrelated with the actual equity markets.
Finally, defensive assets are allowed to receive returns below that of the equity markets so long as they can preserve their value and liquidity in a variety of markets. Treasuries, cash, and low vol strategies all fall into this bucket.