The Endowment Model at 40: What Still Works and What Needs to Change

David Swensen's pioneering work at Yale transformed institutional investing by demonstrating that long-horizon investors could earn significant premiums by accepting illiquidity, complexity, and active management risk. The endowment model—characterized by broad diversification, significant alternative asset allocations, and top-tier manager selection—delivered extraordinary results for decades. Yale's endowment returned 13.7% annualized from 1985 to 2020, far exceeding traditional balanced portfolios. However, the investment landscape has changed dramatically since Swensen first articulated his approach. The proliferation of capital in private markets has compressed returns, manager dispersion has narrowed in some strategies, and the correlation benefits of alternatives have diminished during stress periods. At Harvard, we've been thoughtfully evolving the endowment model while preserving its core insights. The first evolution is in private equity. The industry has grown from $500 billion in assets under management in 2000 to over $8 trillion today. This capital flood has compressed buyout returns—median net IRRs have declined from 15%+ to approximately 11% over the past decade. We've responded by shifting toward co-investments (which reduce fee drag by 150-200 basis points), emerging managers (who show stronger performance persistence), and sector-specialist funds (which demonstrate deeper value creation capabilities). The second evolution concerns hedge funds. The traditional 2-and-20 hedge fund model has been under pressure, with the average hedge fund underperforming a simple 60/40 portfolio over the past decade. We've dramatically reduced our hedge fund allocation from 18% to 8%, concentrating in strategies that provide genuine diversification: systematic macro, volatility arbitrage, and event-driven credit. The key criterion is not absolute return but correlation benefit—we want strategies that perform well when our equity and credit portfolios struggle. The third evolution is in real assets. Traditional real estate and natural resources allocations have been supplemented with infrastructure investments, particularly in digital infrastructure (data centers, fiber networks) and energy transition assets (renewable generation, battery storage). These investments offer inflation protection, stable cash flows, and exposure to secular growth themes. Perhaps the most significant evolution is in our approach to China. The endowment model historically favored emerging market exposure, with China representing a significant allocation. Geopolitical risks, regulatory uncertainty, and governance concerns have led us to reduce direct China exposure while maintaining broader emerging market diversification through India, Southeast Asia, and Latin America. The endowment model's core insight—that long-horizon investors should accept illiquidity and complexity premiums—remains valid. But the implementation must evolve with market conditions. The next generation of the endowment model will be characterized by greater fee sensitivity, more sophisticated risk management, and a willingness to concentrate in areas of genuine competitive advantage rather than diversifying across every alternative asset class.