In this article
← All articles

UNC's 37.8% endowment year: SpaceX and the record versus MSCI World


Executive summary

The UNC Investment Fund returned 37.8% in the year to 30 June 2026, lifted by SpaceX. Over 19 fiscal years from July 2007, the fund compounded at 9.26% a year against 7.97% for MSCI World net total return in US dollars, by our calculation. UNC's figures come from UNC Management Company's report to the UNC-Chapel Hill Board of Trustees (committee book, 23 September 2026, PDF p. 89).

The qualification is where that lead came from. All of it was built in fiscal 2008-2009, when the fund reported a 13.2% loss against 37.0% for the index. From July 2009 to June 2026 the fund compounded at 11.32%, against 11.95% for MSCI World, even with the SpaceX gain. Copio's reading: smaller reported crash losses made the difference to the full-period lead; that does not establish better crash protection or faster future compounding.

What is in the UNC Investment Fund?

The UNC Investment Fund, LLC pools the long-term assets of the University of North Carolina System, its campuses and their foundations. Its reported market value was $17.0bn at 30 June 2026, up from $12.2bn a year earlier (September 2025 committee book, PDF p. 19). UNC-Chapel Hill's own pooled fund is its controlling member with $8.2bn, alongside UNC Health, NC State and other campuses (PDF pp. 80-81).

The latest audited breakdown we could verify is dated 30 June 2025, before SpaceX listed. It appears in UNC-Chapel Hill's financial statements, audited by the North Carolina State Auditor, and covers $12.6bn of investments (audit report, PDF p. 52). That table does not measure fund-level leverage or give financing terms.

UNC Investment Fund investments at 30 June 2025, as a percentage of $12.6bn: private equity partnerships 39.0%, long/short equity hedge funds 15.6%, long-only equity funds 14.0%, real assets partnerships 9.8%, domestic stocks 8.7%, diversifying hedge funds 8.2%, bonds and money market funds 2.2%, credit funds, REITs and other 2.4%.

Holdings at 30 June 2025: private equity partnerships 39.0%; hedge funds and long-only equity funds 38.1%; real assets partnerships 9.8%; domestic stocks 8.7%; bonds and money market funds 2.2%. Source: UNC-Chapel Hill audited financial statements 2024-25, Note 2, PDF pp. 52 and 59; holdings, commitments and redemption terms (CSV). These are audited fair values by investment type, not policy targets.

The two partnership rows, 48.8% of the total, are not redeemable on demand; the notes describe typical ten-year terms with periodic distributions. Hedge funds can be redeemed on terms from 30 days to more than three years. On top of what it held, the fund had promised $1.11bn more to private equity and $0.53bn to real assets, together 13.1% of its investments (PDF p. 59).

Directly held stocks, debt securities, money market funds and listed property trusts came to $1.52bn, about 12% of investments and less than the $1.65bn of unfunded commitments, by our sum. That compares asset and commitment balances, not immediately available cash and bills due on the same day.

The policy targets are different from the holdings. UNC's Strategic Investment Policy Portfolio aims for 27% private equity, 26% long equity, 15% long/short equity, 10% diversifying strategies, 8% fixed income, 8% real estate, 4% energy and natural resources and 2% cash (PDF p. 86). In its FY2024 annual report, UNC put private equity at 32.4% and fixed income at 5.5% as of 30 June 2024 (PDF p. 7).

UNC does not disclose the size of its SpaceX exposure. WRAL reported on 23 September 2026 that the exposure came through venture capital and that payouts from the investment would continue through September 2027. That report does not specify how much will be distributed or whether the payouts will be in cash or shares.

Why does UNC use this structure?

UNC's stated reason is the size of its spending commitment. The fund's stated objective is inflation plus 5.5%, about 8% at the report's long-run inflation assumption of about 2.5% (PDF p. 84). UNC Management Company argues that diversification and private investments are necessary because a traditional stock and bond mix fell short of that objective in its comparison (PDF p. 97).

UNC says a predefined spending rate and a perpetual horizon give it room to forgo liquidity in pursuit of higher long-term returns (PDF p. 87). The fiscal 2024 annual report says the fund favours private investments for their potential illiquidity premium and equities over bonds because of its exposure to inflation (PDF p. 12).

WRAL reported on 23 September 2026 that president Stefan Strein pointed to the initial 2009 venture investment in explaining why the ability to lock up capital can pay. UNC's presentation also names the trade-off: as private capital grows, it has to pay more attention to liquidity elsewhere in the portfolio (PDF p. 87).

Spending is the binding liability. UNC-Chapel Hill's audited statements give a distribution rate of 4.5% of the pool's beginning value in fiscal 2025, within a usual range of 4.0% to 6.0% (PDF p. 14). In fiscal 2026 the Chapel Hill pool paid out $315.1m for programmes and development support, after a net investment gain of $2.30bn (PDF p. 90).

Historical performance versus MSCI World

From June 2007 to June 2026, $100 compounded at the fund's published returns became $537.55. At MSCI World's net USD total returns it became $428.99, and at a 70/30 mix of MSCI World and 10-year US Treasuries $354.99. The fund beat the index in 10 of the 19 fiscal years and had three losing years against five for the index. The paths reinvest reported returns and leave out spending and gifts; the fund's returns are net of fees, and the index carries no implementation costs.

Annual June observations from 2007 show $100 growing to $537.55 at the UNC Investment Fund's returns and $428.99 at MSCI World net USD returns. A second panel shows each fiscal year's return, including fiscal 2009 losses of 19.6% for the fund and 29.50% for the index, fiscal 2022 returns of 4.4% and minus 14.34%, and fiscal 2026 returns of 37.8% and 21.34%.

June 2007-June 2026: $100 becomes $537.55 in the UNC fund and $428.99 in MSCI World; fiscal 2009 cost the fund 19.6% and the index 29.50%; fiscal 2026 added 37.8% against 21.34%. Sources: UNC Investment Fund annual reports for fiscal 2012, 2016, 2021 and 2024, trustees' committee books of September 2025 and September 2026; MSCI World NETR USD; annual returns and wealth (CSV).

Fiscal years 2008-2026UNC Investment FundMSCI World net70/30 World/Treasury
Geometric annual return9.26%7.97%6.90%
Ending value of $100$537.55$428.99$354.99
Sharpe, annual observations0.650.470.56
Sortino, annual observations1.810.941.08
Worst fiscal year-19.6% (2009)-29.50% (2009)-18.56% (2009)
Largest fall, June values-19.6%-37.02%-21.46%

Sharpe and Sortino use the same 19 July-June observations for every column, with US Treasury-bill returns as the cash benchmark and the downside target. They are annual-observation ratios and cannot be compared with ratios built 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, without fund costs. UNC publishes no fund return before fiscal 2008 in any annual report we could retrieve, so the comparison starts in July 2007.

Geometric annual returnUNC fundMSCI World70/30 mix
FY2008-2016 (9 years)4.74%2.50%4.11%
FY2017-2026 (10 years)13.49%13.14%9.46%
FY2008-2025 (18 years, before SpaceX listed)7.86%7.27%6.43%
FY2023-2026 (4 years)14.60%19.06%13.64%

Our trailing returns match UNC's own figures to the published decimal: 20.1% over three years, 12.5% over five, 14.8% over seven, 13.5% over ten and 11.4% over fifteen (PDF pp. 97-98). UNC also reports 9.9% over twenty years; we cannot match that window because no fetched document gives the fiscal 2007 return.

The split shows where the lead came from. In fiscal 2008-2009 the fund lost 13.2% while MSCI World lost 37.0%, by our calculation. From fiscal 2010 to 2025, before SpaceX listed, the fund compounded at 9.85% a year against 11.39% for the index; adding fiscal 2026 brings the post-crash figures to 11.32% and 11.95%. Over the ten years to June 2026 the fund led by 0.35 points a year, and over the four years since June 2022 it trailed by 4.46 points despite the 37.8% year.

The UNC Investment Fund's historical risk-return point against a frontier restricted to annually rebalanced mixes of MSCI World and a modelled 10-year US Treasury, fiscal 2008-2026. The horizontal axis is the standard deviation of annual excess returns; the vertical axis is arithmetic mean return. UNC fund: 13.22% risk, 10.01% mean, 9.26% CAGR, Sharpe 0.65 and Sortino 1.81. MSCI World: 16.58% risk, 9.28% mean, 7.97% CAGR, Sharpe 0.47 and Sortino 0.94. Lagging private valuations can flatter the fund's position.

The listed mix with the closest measured risk, 83% MSCI World and 17% Treasury, had 13.25% risk against the fund's 13.22% and compounded at 7.41% a year against 9.26%. Sources: the same UNC, MSCI, US Treasury and cash inputs; World/Treasury frontier data (CSV). The listed mixes are historical illustrations with annual rebalancing and no implementation costs, not proposed allocations.

The fund compounded faster than the listed mix with the nearest measured volatility, and sits above the curve on arithmetic mean return. It did not have the highest Sharpe ratio: the best listed mix, 39% equities, scored 0.67 against the fund's 0.65. The fund's Sortino ratio of 1.81 is far higher, because it reported only three years below cash, and those losses were smaller than the index's. The chart's listed assets cannot represent UNC's opportunity set or its access to venture managers.

Private valuations are one possible explanation for that smoothness. In fiscal 2022 the fund reported 4.4% while MSCI World lost 14.34%; in fiscal 2023 it reported -0.4% while the index gained 18.51%. Regressing the fund's annual return on the index's same-year and previous-year returns gives coefficients of 0.66 and 0.25 on 19 observations. That pattern is consistent with delayed valuations, but it does not isolate their effect.

Copio's reading

Is it optimal for geometric growth?

UNC's higher reported compound return over 19 years does not establish a growth optimum. Judge expected log wealth after fees over decades, with the fund's ongoing payouts, capital calls and liquidity needs held constant. An all-equity index is a reference point; the reported record does not show how a liquid stock and bond mix would fare under the same obligations and financing constraints.

First, concentration. WRAL reported on 23 September 2026 that Strein attributed about half of the year's growth to SpaceX, without defining growth; UNC's written slides say the fund ranked in the top decile even without SpaceX, with a threshold of 20.0%. In a separate sensitivity scenario, halving the fund's FY2026 return to 18.9% shrinks the 19-year lead from 1.29 points a year to 0.44 and puts its ten-year CAGR at 11.83%, below MSCI World's 13.14%.

The scenario tests how sensitive the record is to one exceptional year, not what the fund earned without SpaceX. For a household, recalculate compound growth without the largest contribution in a strong year to see how much the result depended on one winner.

Second, protection. The fund's best contribution to compound growth came from losing less in 2008-2009, which left more capital to compound afterwards. Diversification and lagged pricing could both explain the cushion; the annual data cannot separate them. From fiscal 2010 to 2025 the fund compounded 1.54 points a year below the index, but its mix changed and the return gap does not measure the cost of protection.

Third, costs and cash flows. Returns are net of all fees, so manager fees are already inside the 9.26%. UNC Management Company charges members 0.35% on the first $50m and 0.20% above that, plus about 0.02% of administrative costs (FY2024 report, PDF p. 36). The external managers' fees, charged on the private partnerships, hedge funds and long-only funds that made up nearly 87% of investments in 2025, are not disclosed in the documents we fetched.

The decisive comparison is with a feasible alternative after all costs, cash flows and the need to fund capital calls in bad markets. The record shows that UNC's smaller reported losses in 2008-2009 mattered more to the long-run ranking than its ability to beat the index year after year.

Is it crash aware and protected?

UNC plans for liquidity and smooths spending; its reported crash record is better than the index, but smoothed marks and a payout cut mean the protection was partial. The worst reported fiscal year, -19.6% in 2009, compares with -29.50% for MSCI World. UNC's own chart puts the fund's loss at 16.5% while the S&P 500 fell 50.9% in 2007-2009, and at 7.2% against 23.9% in the first nine months of 2022 (FY2024 report, PDF p. 35).

Those comparisons use a US stock index, not a portfolio with UNC's assets and spending obligations. Fiscal 2023's -0.4% in a rising equity market is consistent with delayed valuations, not proof that losses were merely deferred. Year-end observations also cannot show the fund's worst intra-year fall.

Spending still fell after the loss. Chapel Hill's payout per unit fell from $418 in fiscal 2009 to $387 in fiscal 2010, a 7.4% cut, and did not pass the 2009 level until fiscal 2014 at $420 (committee book, September 2025, PDF p. 41). The university's endowment policy tells departments in severe declines to examine "possible deferment of program expenses", with invasion of principal "an option of last resort" (audit report, PDF p. 14).

The June 2025 balances do not reveal when commitments will be called or what assets could be sold in a stressed market. Some hedge funds were redeemable but with notice periods and restrictions. The household version is to compare payments coming due with assets that are accessible in a downturn, rather than treating every account balance as available cash.

Is it diversified without giving up growth?

The fund spreads money across managers and vehicles, but its reported annual returns move closely with world equities, and one company materially boosted fiscal 2026. On annual observations, the fund's return correlated 0.78 with MSCI World, 0.18 with the index's previous-year return and 0.91 with its own policy benchmark. The 10-year Treasury correlated -0.36 with the fund and -0.41 with the index, but bonds and money market funds were only 2.2% of the fund's investments in 2025.

Correlation matrix of annual July-June returns, fiscal 2008-2026. UNC Investment Fund with MSCI World 0.78; with the previous year's MSCI World return 0.18; with the modelled 10-year Treasury minus 0.36; with its own policy benchmark 0.91; with the 70/30 mix 0.75. MSCI World with the Treasury minus 0.41. Observed annual correlations, not crash correlations.

Observed correlation of annual returns, FY2008-2026: UNC fund with MSCI World 0.78; with the prior year's MSCI World 0.18; with the 10-year Treasury -0.36; with its policy benchmark 0.91. Source: UNC's published returns, MSCI World NETR USD and US Treasury yields; correlation matrix (CSV). These are observed annual correlations on 19 points, not crash correlations or UNC's modelling assumptions.

Leaving out fiscal 2026, the correlation with MSCI World rises to 0.81. Adding the two regression coefficients gives about 0.90, a measure of equity sensitivity in this model, not an estimate of the fund's equity weight or its response in a crash.

The fund's equity-heavy mix lost less on reported values in 2009, but the record does not establish an independent return source that reliably holds up in a joint decline. Copio does not grade the fund in the portfolio checker, because nearly 87% of it sits in private partnerships and externally managed hedge and equity funds that no UCITS ETF replica can represent faithfully.

Our studies of Yale's fiscal 2026 and Harvard's endowment apply the same questions to two other private-asset pools. Scottish Mortgage's SpaceX position shows the same company dominating a listed portfolio, and our Kelly criterion study explains why position size, not only the payoff, decides long-run growth.


Sources and methodology

Checked on 20 November 2026. Fiscal 2026 results are dated 30 June 2026 and come from "UNC Investment Fund, LLC Fiscal Year 2026 Review", presented on 23 September 2026 in the Budget, Finance and Infrastructure Committee book: entity structure and market values (PDF pp. 79-81), return objective (p. 84), policy targets (p. 86), time horizon and liquidity (p. 87), results (p. 89), Chapel Hill pool flows (p. 90), long-term returns (pp. 97-98) and wrap-up (pp. 101-102).

Fiscal 2025 comes from the September 2025 committee book, PDF pp. 19, 28 and 41 (market value, return and payout per unit). Holdings, unfunded commitments, redemption terms, the distribution rate and the endowment policy come from the State Auditor's audit report on UNC-Chapel Hill for the year to 30 June 2025, dated 13 November 2025, PDF pp. 14, 52 and 59.

Annual returns for fiscal 2008-2024 come from the "Fund Return" rows of UNC Investment Fund annual reports: 2012, 2013, 2014 and 2016, archived copies of UNC Management Company's own files, and 2019, 2020, 2021, 2022, 2023 and 2024. UNC returns with locators (CSV) also list the policy benchmark, which the fiscal 2021 and 2022 reports show as 32.6% for fiscal 2021 and later reports as 32.7%; we use 32.7%.

Press: WRAL, 23 September 2026, for the "about half" attribution, the initial 2009 venture investment and the payout schedule to September 2027, none of which appear in UNC's written report.

The equity benchmark is MSCI World net total return in USD, index 990100, from MSCI's month-end levels. 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 2008-2026, with the index's fiscal 2007 return as the first previous-year value.

The sensitivity scenario replaces the fiscal 2026 return with 18.9%, half of 37.8%, and recomputes the 19-year and ten-year CAGRs. Holdings shares divide each audited investment type by $12,600,956,226; the direct-holdings sum adds domestic and foreign stocks, REITs and debt securities, including money market funds. Unfunded commitments add the two partnership rows in Note 2.

The fund's returns are net of fees and rely on managers' estimates for private funds, which lag listed prices; UNC-Chapel Hill's statements note that the fund's returns were calculated before the annual fair-value adjustment. The index omits implementation costs; neither side includes spending, gifts, household taxes or inflation. Annual observations hide intra-year falls, and the fund's mix changed over 19 years, so this is actual fund performance, not a backtest of today's holdings. This article is educational, not personalised investment advice.

FAQ

What return did the UNC endowment fund earn in fiscal 2026?

The UNC Investment Fund, which pools the long-term assets of the University of North Carolina System, returned 37.8% net of fees in the year to 30 June 2026 and was worth $17.0bn. Its policy benchmark returned 16.5%, a 70/30 global stock and bond portfolio 17.5% and the median of BNY's endowment and foundation universe, adjusted for internal fees, 15.0%, according to UNC Management Company's report to UNC-Chapel Hill's trustees on 23 September 2026.

How much of UNC's return came from SpaceX?

UNC's written report names SpaceX as a major contributor but gives no figure. WRAL reported on 23 September 2026 that UNC Management Company's president, Stefan Strein, told a trustees committee that about half of the fund's growth was attributable to SpaceX. The report also says the fund ranked in the top decile of its peer group with and without SpaceX; the top-decile threshold was 20.0%.

Has the UNC Investment Fund beaten MSCI World?

Over the 19 fiscal years from July 2007 to June 2026, the fund compounded at 9.26% a year against 7.97% for MSCI World net total return in US dollars, by our calculation from UNC's published annual returns. The whole lead came from fiscal 2008-2009, when the fund lost 13.2% and the index 37.0%. From July 2009 to June 2026 the fund compounded at 11.32% against 11.95% for the index, including the SpaceX year.

What does the UNC Investment Fund invest in?

At 30 June 2025, the latest audited breakdown we could verify, 39.0% of its $12.6bn of investments sat in private equity partnerships and 9.8% in real assets partnerships, neither redeemable on demand. Hedge funds and long-only equity funds made up 38.1%, and directly held stocks, bonds, money market funds and listed property about 12%. The fund had $1.65bn of unfunded commitments to private partnerships.

Can a household copy the UNC endowment's SpaceX return?

A household cannot reproduce the reported SpaceX outcome from the fund's disclosures. WRAL reported on 23 September 2026 that UNC made its initial investment through a venture capital fund in 2009. A household can test its own dependence on one winner by comparing its long-run result with a broad index after removing that holding's contribution in a strong year.

Optional cookies help us measure visits and advertising. Choose each purpose, or decline both. You can change your choice using Cookie settings in the footer. Cookie policy.