Mike Hennessy defends the endowment model

September 14, 2009  

The endowment model is very difficult and internally expensive to implement well. It is clearly worth the candle, but no less intimidating

Michael Hennessy, co-founder and managing director of investments at $9 billion Morgan Creek Capital Management, takes issue with popular reporting about endowments. Sure, a lot of them have lost money in 2008, but over any reasonable period of time, as he says below, they have outperformed other models of investing. Reprinted below is Hennessy's "Rebuttal to the Endowment Model Criticism in the Press," which he wrote for clients of Morgan Creek.

There has been much written in the popular press lately about the failures and even the "death" of the endowment model. The discourse regarding this matter has been surprisingly simplistic, naive and exceedingly short sighted. As was the case with Mark Twain, reports on the death of the endowment model have been greatly exaggerated. Let's start with the facts. The "endowment model" practiced by most of the big university endowments and many big foundations (but also by some astute smaller endowments and foundations) has overwhelmingly outperformed virtually all other models over any reasonable time period, and has done so for a very long time now. There is no single model, mode or manner of investing that outperforms in every environment and over every time period, and the endowment model of investing was never predicated on being the exception to this obvious reality. In fact, endowments' time horizons are as long as any investor's horizon, and hence are strictly focused on the long term. This is a huge advantage because there is clearly a significant liquidity premium to be captured by investing long term, not to mention the ability to better avoid the chaotic noise and behavioral finance mistakes that arise with a short term environment and outlook - especially in volatile markets.

Interestingly though, the endowment model has generally outperformed in most environments (bull/bear, value/growth, small cap/large cap, domestic/international) as well as over most shorter time periods. Additionally, the model typically has done so with less risk. In fact, in the "failure" of 2008 - by any measure a sorely disappointing year - most of the big endowments still performed no worse than, and in many cases better than, many institutional and retail investors, as well as the 70% S&P / 30% Barclays Gov't/Corp Bond benchmark, and the MSCI World Index. How has it been able to accomplish these challenging feats?

The endowment model is very difficult and internally expensive to implement well. It is clearly worth the candle, but no less intimidating. Experienced, bright, diligent and successful staff must be attracted, properly incentivized, and retained, as well as given sufficient resources to implement the model. The model employs the broadest asset allocation possible, literally encompassing all asset classes globally and virtually all strategies globally. There is no more expansive landscape in which to invest. This leads to a maximum amount of diversification, but in a way that does not over-diversify (or "diworsify") if properly implemented. Because there are ranges around the targets for each asset class (and each strategy), the model also allows an ample amount of flexibility to underweight and overweight asset classes and strategies. This flexibility is constrained, however, by the allowable floors and ceilings that are set for these asset classes and strategies. So the model by definition is a highly disciplined process. It effectively promotes rebalancing, "buying low" and "selling high." When done skillfully, great value can be added. Lastly, the big endowments have shown the ability to identify, evaluate, gain access and allocate to highly skilled investment managers to implement their mandates. Particularly in the world of so called "alternative investments," the performance dispersion between a top quartile manager and an average manager is very wide. Skill at asset allocation, portfolio construction and manager selection can add, and typically has added, great value not only in investment returns but in risk management as well.

So, what happened in 2008? The greatest financial crisis in anyone's working lifetime, of course, where all correlations went to +1.0. In fact, some of the best assets performed worst because they were the only assets that transacted during this severest of liquidity crunches. The rare liquidity provider (i.e., buyer) would typically only buy the best assets, and had the bargaining power to squeeze prices far below anyrational value. Meanwhile, many bad assets tended to "hold their own" better because they just didn't transact at all. Correlations always rise when crises occur, and liquidity gets crunched most in a crisis. It was a lack of liquidity which hurt endowments the most. Some (not all) endowments were far too aggressive with their private assets programs, sometimes to the point of planning on incoming charitable contributions (and even seasonal tuition payments) to help fund private asset capital commitments and private capital calls in a "just-in-time" fashion. But in a severe crisis, donations dry up, liquidity dries up, fear and panic ensue, etc. ad nauseam. Exacerbating this illiquidity were hedge funds that "gated" investments or suspended redemptions altogether, making it more difficult, if not impossible, to garner sufficient liquidity that otherwise would have been available in a properly designed hedge fund program with a "steady-state" mode of available liquidity.

Having worked most of my career on the senior investment staffs at the endowments of Duke University and the University of North Carolina, and currently serving on the investment committee of the College of William & Mary Foundation, I am surprised that some practitioners seem to have cast aside an essential precept that has always been integral to the endowment model: have a sufficient amount of liquidity available! Endowments and foundations usually have a highly defined liability stream (typically clustering around 5% annually) as well as a variety of means to smooth even that liability stream over time. It is thus imperative, and not especially exigent, to dependably have this liquidity when due. While the most reliably liquid assets tend to be a drag on long term performance, this drag can be more than offset by skillful implementation of the rest of the model. The "dry powder" value of liquidity then rises to its highest level when it naturally is most desired and needed, e.g., in the midst of crisis -- continuing even today when liquidity premiums are exceptionally wide. One relatively new twist on the model is a result of the industry having evolved to make available a staggering array of efficient, inexpensive and highly liquid investment vehicles such as ETFs (Exchange Traded Funds) and other liquid instruments which make it easier, more effective and less punitive than ever to infuse liquidity within the model.

However, more importantly for this discussion now is: what will work best in the future? Stocks? Stocks don't appear to be highly undervalued nor are they expected to lower risk. Government bonds? Given the massive government stimulus, these bonds might not be the best course. Cash? Guaranteed to not deliver your real return hurdle. Market timing? No one has ever been able to do so proficiently and consistently. Given the major dislocations that the great bubble has caused, the wide liquidity premiums available, the continuing rationalization of the "alternative investment" arena, and the evolution of the industry as a whole, it would seem premature to be trumpeting the death of the Endowment Model just yet. My partners and I at Morgan Creek (as well as many other experienced and successful peers and competitors) are putting our money where this writer's mouth is, and we are confident that the endowment model is as well poised for the future as any model could be.

Some previous AR coverage of Hennessy:
In Pursuit of Endowment-Style Investing
The sweep-up begins


The endowment model as practiced by most of the biggest endowents is a flexible, opportunistic but highly disciplined and highly diversified long term investment strategy designed to preserve and grow the purchasing power of endowment in perpetuity (i.e., a return objective of at least spending rate + inflation). The model does reflect many of the investing tenets espoused and practiced by Yale's David Swenson, but is customized to a university's specific spending rate, reliance on endowment, risk tolerance, needs for liquidity, and resources dedicated to the endeavor. As such it generally does have a relatively heavy allocation to "alternative assets." Most of the big endowments have been practicing this model for a good two decades, almost equivalent to Swenson's tenure at Yale. That period alone reflects ample data points and different market cycles. But again, the question that counts is: what is the best model going forward? Jeff Tarrant and Ted Seides of Protege Partners have made a famous bet with Warren Buffet (albeit just hedge funds vs. the S&P over the next 10 years). Not only am I siding with Protege, but it probably won't even be close. I predict the full endowment model does better still.

Mike Hennessy Sep 19, 2009

How is Henessy defining "the endowment model?" If most large endowments have been trying to emulate Swensen's (Yale's) uber-alternative investment strategies does it really make sense to roll in track records from the pre-Swenson era when endowments invested almost exclusively in liquid, income generating securities?

Brendan Pollock Sep 15, 2009

Bank of America/Merrill Lynch provides the innovative research and market insights that will help you succeed, including commentary on the key themes for the year ahead as well as trends driving growth in important industries and sectors. You will gain access to global research and the other information you need to make informed investment decisions.

Click here for more details.

Latest Poll

Do activist hedge fund investors ultimately add value to the companies they target?

 - 59%
 - 41%

View previous results