Yale's endowment returned 26.3% in the year to 30 June 2026 and, over the 25 fiscal years since July 2001, compounded at 10.78% a year against 8.10% for MSCI World net total return in US dollars. A 70/30 mix of MSCI World and 10-year US Treasuries compounded at 7.12%. Yale's figures come from its own audited financial reports; the comparison is our calculation on matched July-June years (Yale News, 27 October 2026).
The qualification is the timing. Yale built its whole lead over the index between fiscal 2002 and 2016, when it compounded at 10.21% against 4.86%. Over the ten years to June 2026 it compounded at 11.65% against 13.14%, still ahead of the 70/30 mix at 9.46%. Copio's reading: the record shows higher reported compound returns over 25 years, with a weaker last decade against world equities. Access to private managers and lagged valuations complicate the comparison with a listed portfolio.
Yale has not published asset-class weights since the fiscal 2021 targets covered in our Swensen study, but its audited accounts value its investments by asset class. At 30 June 2026, the university held $59.3bn of investments, which include amounts invested for affiliated institutions and are stated before $3.6bn of investment-related liabilities. The endowment itself accounted for about $53.3bn of net assets (Financial Report 2025-2026, PDF pp. 18 and 37).
Holdings at 30 June 2026: venture capital 25.9%; developed-market equity funds 20.5%; leveraged buyouts 19.4%; marketable alternatives 13.9%; real assets 9.4%; cash, bonds and directly held stocks 6.8%. Source: Yale Financial Report 2025-2026, Note 2, PDF pp. 37-41; holdings and commitments (CSV). These are audited fair values, mostly external managers' own estimates of net asset value, not policy targets.
Leveraged buyouts and real assets, 28.7% of investments, sit in closed-end funds that Yale's accounts list as "not eligible for redemption". Venture capital ranges from semi-annual redemption to closed-end structures. On top of what it holds, Yale had promised $11.2bn more to its funds: $5.2bn to buyouts, $2.5bn to venture capital and $2.5bn to real assets (PDF p. 41).
The mix moved further into private equity during fiscal 2026. Venture capital holdings rose from $11.3bn to $15.4bn, while cash and short-term investments fell from $2.9bn to $1.7bn and fixed income from $2.2bn to $1.3bn (PDF pp. 37-38). The accounts also disclose $279m of securities sold short and derivatives, so the asset shares alone are not a complete net-exposure map. Yale's report describes roughly 97% of the endowment pool as invested in assets expected to produce equity-like returns (PDF p. 19).
Yale's stated reasons are spending, inflation and time. Its 2025-2026 report says that funding current operations while preserving purchasing power "dictate investing for high returns, which leads the endowment to be weighted toward equities", and that vulnerability to inflation "directs the university away from fixed income instruments" (PDF p. 19). The 2026 release describes an endowment structured "in perpetuity".
The case for private funds is older and more specific. Yale's 2015 endowment report says it sought "approximately one-half of the portfolio" in illiquid buyouts, venture capital, real estate and natural resources, arguing that alternative assets "tend to be less efficiently priced" and that its "long time horizon is well suited to exploit illiquid, less efficient markets" (The Yale Endowment 2015, PDF p. 9).
The same report concedes that its quantitative tools "have difficulty incorporating factors such as market liquidity or the influence of significant, low-probability events" (PDF p. 8). Yale's Investments Office describes the model as "equity-oriented and diversified across asset classes", built on partnerships with external managers whose identities it keeps confidential (The Endowment FAQ).
Spending is the binding liability: Yale targets spending of 5.25% a year; each year's amount is 80% of the previous year's spending plus 20% of the target rate applied to the endowment's value at the start of the prior year, adjusted for inflation and kept between 4.0% and 6.5% (PDF p. 18). Over the decade to June 2026, Yale spent $16.3bn, equal to 64% of the endowment's value at the start of that decade (Yale News).
Two newer pressures sit outside the original model. The federal tax on Yale's net investment income rose from 1.4% to 8% on 1 July 2026, which Yale's provost estimates will cost about $350m a year (Fall 2026 Financial Update). And Yale has been selling private-fund stakes: its audited accounts record a sale completed on 30 June 2025, with a $1.04bn receivable collected in fiscal 2026 (PDF p. 38).
Yale gives no reason or size for the sales. Secondaries Investor reported on 25 September 2026 that HarbourVest had bought into a roughly $1bn Yale portfolio sale, following a 2025 deal of $3-4bn "with pricing in the mid-90% of NAV" (Secondaries Investor). We could not confirm those sizes from Yale. If accurate, Yale accepted roughly a 5% discount to reported value to turn private stakes into cash.
From June 2001 to June 2026, $100 compounded at Yale's published returns became $1,293.72. At MSCI World's net USD total returns it became $700.26. Yale beat the index in 15 of the 25 fiscal years and had one negative year, fiscal 2009, against seven for the index. The wealth paths reinvest the reported returns and omit spending and gifts; Yale's returns are net of fees, and the index carries no implementation costs.
June 2001-June 2026: $100 becomes $1,293.72 at Yale and $700.26 at MSCI World; fiscal 2009 cost Yale 24.6% and the index 29.50%. Sources: Yale Financial Reports 2002-03, 2006-07, 2011-12, 2016-17, 2021-22 and 2025-26, five-year overview tables; MSCI World NETR USD; annual returns and wealth (CSV).
| Fiscal years 2002-2026 | Yale endowment | MSCI World net | 70/30 World/Treasury |
|---|---|---|---|
| Geometric annual return | 10.78% | 8.10% | 7.12% |
| Ending value of $100 | $1,293.72 | $700.26 | $558.79 |
| Sharpe, annual observations | 0.79 | 0.48 | 0.58 |
| Sortino, annual observations | 1.93 | 0.94 | 1.13 |
| Worst fiscal year | -24.6% (2009) | -29.50% (2009) | -18.56% (2009) |
| Largest fall, June values | -24.6% | -37.02% | -21.46% |
Sharpe and Sortino use the same 25 July-June observations for every column, with US Treasury-bill returns as the cash benchmark and the downside target. They are annual-observation ratios, not comparable with ratios computed from monthly data. The 70/30 mix is rebalanced each July and uses a 10-year Treasury return modelled from the US Treasury's month-end yields, so it carries no fund costs. The split by period matters more than the 25-year total.
| Geometric annual return | Yale | MSCI World | 70/30 mix |
|---|---|---|---|
| FY2002-2016 (15 years) | 10.21% | 4.86% | 5.59% |
| FY2017-2026 (10 years) | 11.65% | 13.14% | 9.46% |
| FY2022-2026 (5 years) | 8.76% | 11.47% | 7.63% |
Our ten-year Yale figure rounds to the 11.7% Yale reports, which checks the input series. Yale says it beat "a typical 70/30 stock and bond portfolio" by 2.7 points a year over that decade; our listed 70/30 gives 2.19 points, and Yale does not define its version. Against a global equity index, Yale's lead belongs to the first fifteen years.
The closest tested mix, 82% MSCI World / 18% Treasury, had 12.26% measured risk against Yale's 12.35%, and compounded at 7.56% a year against 10.78%. Sources: the same Yale, MSCI, US Treasury and cash inputs; World/Treasury frontier data (CSV). These are historical comparisons with annual rebalancing and no costs, not proposed allocations.
Yale sits well above every listed stock/bond mix on this map. The best Sharpe ratio among them was 0.73, for 38% equities, against Yale's 0.79. The compound-return lead is measurable, but the apparent risk advantage also depends on private-fund estimates that can lag listed prices and lower measured volatility. The chart's listed assets cannot represent Yale's full opportunity set or its access to external managers.
The lag shows in the record. In fiscal 2022, Yale reported 0.8% while MSCI World lost 14.34%. Over fiscal 2022-2024 together, Yale gained 8.5% against 22.0% for the index, by our calculation. Regressing Yale's annual return on the index's same-year and previous-year returns gives coefficients of 0.68 and 0.18; the lagged term is weakly significant on 24 observations.
Yale's 25-year reported compound return exceeds the listed comparators, but does not establish a growth optimum. The objective here is expected log growth of the endowment over a perpetual horizon, after fees and tax, while paying about 5% a year and meeting capital calls. The historical return calculation favours Yale over 25 years, but excludes the spending and tax in that objective. Three tests decide whether that result carries forward.
First, concentration: about 97% of the pool is equity-like and 45.3% of investments sit in venture capital and buyouts. That concentration was paid for between 2002 and 2016, but not against world equities over the last decade, when Yale trailed by 1.49 points a year. Yale credits the decade's result with $14.3bn of value above its composite passive benchmark (PDF p. 18), which is consistent with a contribution from manager selection. That comparison does not isolate manager skill from asset mix, benchmark design or valuation effects.
Second, protection: bonds were 2.2% of disclosed university investments, while equity swaps and options had fair values of $190m in assets and $205m in liabilities (PDF p. 41). Fair value does not reveal their notional exposure, option strikes or crash payouts. The accounts therefore cannot establish either the extent or the cost of any portfolio hedge.
Third, costs and cash flows: fees are already inside the returns, while the tax rate rose from 1.4% to 8%. Yale's estimated annual tax bill of $350m equals about 0.64% of the $54.5bn endowment, rather than a measured incremental return reduction. Reported discounts on secondary sales suggest another possible liquidity cost.
A feasible alternative must meet the same spending and capital-call constraints, so an all-equity index is not a like-for-like replacement. The decisive comparison still missing is Yale's expected manager excess after the new tax, against a listed portfolio with the same spending rule. Copio's verdict: stronger reported compound returns over the full period, dependence on private-manager access, and a ten-year return below world equities after 2016.
Spending and commitment controls help liquidity; the record does not establish crash protection. Yale's worst reported fiscal year, -24.6% in 2009, was smaller than the index's -29.50%, and over fiscal 2008-2009 Yale lost 21.2% against 37.0% by our calculation. Annual observations and lagged private valuations can hide losses in the worst months; these figures do not measure the maximum economic drawdown.
The liquidity test came in 2009. The endowment fell 29.1% to $16.3bn after returns and spending; Yale reduced expenses, deferred much of its building programme and prepared its fiscal 2010 budget with significantly less endowment income, including some elimination of positions (Financial Report 2008-2009, PDF p. 4). At 30 June 2009 it had $7.6bn of uncalled commitments, while management estimated it could liquidate about $3.6bn (unaudited) for short-term needs (PDF p. 39).
The position today is less stretched: uncalled commitments equalled 47% of endowment net assets in June 2009; in June 2026, $11.2bn equalled 21% of about $53.3bn, by our division. That is the evidence of crash awareness: smaller promises relative to the pool, and a spending rule that only moves 20% of the way towards a lower endowment value each year. Neither the smaller commitment ratio nor the spending rule establishes a payoff when equities fall; the published derivative values cannot resolve that question.
The confirmed 2025 sale shows Yale converted investments into a receivable, collected the following year; the press reports suggest discounts can make that liquidity costly. Yale does not disclose the reason for that sale. Our crash study of asset classes compares listed assets over several falls covered by this return record.
Yale spreads assets across managers and vehicles, but its annual returns remain strongly linked to world equities. On annual observations, Yale's returns correlated 0.83 with MSCI World and 0.17 with the index's previous-year return. The 10-year Treasury correlated -0.44 with the index and rose in six of its seven losing years, fiscal 2022 being the exception; Yale's disclosed fixed-income allocation is small.
Observed correlation of annual returns, FY2003-2026: Yale with MSCI World 0.83; with the prior year's MSCI World 0.17; with the 10-year Treasury -0.27; MSCI World with the Treasury -0.44. Source: Yale's published returns, MSCI World NETR USD and US Treasury yields; correlation matrix (CSV). The sample starts in fiscal 2003 because the index series we use starts in December 2000; these are observed annual correlations, not crash correlations.
The 0.83 correlation is consistent with Yale's stated equity orientation; the positive lag coefficient is consistent with delayed pricing. Neither result isolates the behaviour of each private asset class. The record supports diversification within an equity-heavy strategy, but does not establish reliable protection against a joint fall in public and private assets. Copio does not grade Yale in the portfolio checker, because most of its assets are private funds that no listed ETF replica can represent faithfully.
Yale's record combines confidential external managers, a perpetual horizon and the capacity to meet capital calls for years. An ETF portfolio cannot faithfully reproduce that combination. Yale's spending rule separates withdrawals from the latest market value; a household can write down its own withdrawal rule before a fall.
Yale's 2009 squeeze came from commitments larger than its accessible cash. The household version is to count locked money, such as a pension, property or a private fund, separately from what can be sold within a week, and compare the second figure with known payments. Finally, compare any fund's result with a listed index over the same dates, as this article does, before crediting a smoother path with protection.
Our studies of Harvard's endowment and CalPERS apply the same questions to other private-asset-heavy institutions; CalPERS compounded 4.56 points a year below MSCI World over its matched decade. For why the Sharpe ratio and compounded growth can point in different directions, see Sharpe ratio versus CAGR.
Checked on 13 November 2026. Holdings, commitments and the spending rule are dated 30 June 2026. The Yale Financial Report 2025-2026, with statements dated 27 October 2026, supplies the five-year overview (PDF p. 3), endowment, spending and allocation text (pp. 18-19), investments by asset class (pp. 37-38), the 2025 sale receivable (p. 38) and redemption terms and unfunded commitments (p. 41).
Annual returns for fiscal 2002-2026 come from the "Total return on investments" row of the five-year overviews in the reports for 2002-03, 2006-07, 2011-12, 2016-17, 2021-22 and 2025-26, listed on Yale's audited financial reports page. Yale's returns with locators (CSV) also list fiscal 1999-2001, which fall outside the comparison.
Other Yale sources: Yale News, 27 October 2026 (fiscal 2026 return, distributions, ten-year comparisons, decade spending); the Financial Report 2008-2009, PDF pp. 4 and 39 (2009 budget response, commitments and liquidity estimate); The Yale Endowment 2015, a copy on a Yale course server, PDF pp. 8-9; the Investments Office FAQ; and the provost's Fall 2026 Financial Update, 27 October 2026.
Press: Secondaries Investor, 25 September 2026, headline and summary only, unconfirmed by Yale. Bloomberg Government published "Yale's Endowment Model Falters at Tough Moment for Universities" on 26 January 2026, before the fiscal 2026 result; we could not read the full text and use none of its figures.
The equity benchmark is MSCI World net total return in USD, index 990100, from MSCI's month-end levels, which start on 29 December 2000. Cash is the RF column of Kenneth French's monthly factors, compounded July-June. The Treasury leg uses month-end 10-year yields from the US Treasury par yield curve. Monthly inputs (CSV) support reproduction.
The Treasury return is modelled, not a fund: each month a par bond is bought at the previous month-end yield and repriced at the new yield with 119 months left, plus one month of coupon. The 70/30 mix holds 70% MSCI World and 30% of that bond, rebalanced each 1 July, without costs or tax. CAGR is the product of annual growth factors to the power of one over the number of years, minus one.
Sharpe is mean annual excess return over cash divided by its sample standard deviation. Sortino divides the same mean by the square root of the mean squared negative excess return, counting all years. There is no annualisation multiplier. The frontier covers 101 World/Treasury mixes; its risk axis is the standard deviation of annual excess returns. Correlations and the lag regression use fiscal 2003-2026, because the index's fiscal 2001 return is unavailable.
The holdings shares divide each asset class's audited fair value by total investments of $59,257.9m; fund rows combine net-asset-value and Level 3 amounts as in Yale's redemption-terms table. Commitment ratios divide unfunded commitments by endowment net assets ($7.6bn / $16.3bn in 2009; $11.2bn / $53.3bn in 2026), and the tax ratio divides $350m by $54.5bn; all are our calculations.
Yale's returns are net of fees and rely on manager-estimated values for private funds, which lag listed prices. The index omits implementation costs; neither side includes spending, gifts, household taxes or inflation. Annual observations hide intra-year falls, and Yale's allocation changed over the 25 years, so this is actual fund performance, not a backtest of today's holdings. This article is educational, not personalised investment advice.
Yale reported a 26.3% return, net of fees, for the year to 30 June 2026, an investment gain of $11.3bn. After $2.2bn of distributions to the operating budget, the endowment was valued at $54.5bn. Yale's audited financial report puts endowment net assets at about $53.3bn on the same date, a different accounting measure.
Over the 25 fiscal years from July 2001 to June 2026, Yale compounded at 10.78% a year against 8.10% for MSCI World net total return in US dollars, by our calculation from Yale's published annual returns. The whole margin came in fiscal 2002-2016. Over the ten years to June 2026, Yale compounded at 11.65% against 13.14% for the index.
Yale's audited accounts list $59.3bn of university investments at 30 June 2026: 25.9% venture capital, 20.5% developed-market equity funds, 19.4% leveraged buyouts, 13.9% marketable alternatives and 9.4% real assets. Cash, bonds and directly held shares made up 6.8%. Yale also had $11.2bn of unfunded commitments to private funds.
Its reported losses were smaller than the index's, but that is not the same as protection. Yale lost 24.6% in fiscal 2009, against 29.50% for MSCI World. In June 2009 it faced $7.6bn of uncalled commitments while estimating it could liquidate about $3.6bn, and it reduced expenses. Lagging private valuations also make reported losses look smaller.
Not its return source. Yale's record rests on access to external private-fund managers, a perpetual horizon and the capacity to meet capital calls for years. A household can copy narrower disciplines: a written withdrawal rule, illiquid commitments sized against cash that is actually available, and performance checked at matched dates against a listed benchmark.
